OFCOM's guidance giving examples of content that is harmful to children

Protecting Children from Harms Online

Guidance on Content Harmful to Children¶

Published 24 April 2025

Contents¶

Section¶

1. Introduction to Ofcom’s Guidance on Content Harmful to Children ................................ 3

2. Guidance on pornographic content ................................................................................. 13

3. Guidance on suicide content ............................................................................................ 20

4. Guidance on self-harm content ....................................................................................... 26

5. Guidance on eating disorder content .............................................................................. 32

6. Guidance on abuse and hate content .............................................................................. 41

7. Guidance on bullying content .......................................................................................... 47

8. Guidance on violent content ............................................................................................ 53

9. Guidance on harmful substances content ....................................................................... 62

10. Guidance on dangerous stunts and challenges content .................................................. 67

Section 1 Introduction to Ofcom’s Guidance on Content Harmful to Children¶

Content warning: this guidance contains detailed descriptions of harmful content. This document discusses topics that may be very upsetting. We have endeavoured to treat such topics sensitively and note where a proper reflection of content that is harmful to children requires us to include themes which may cause distress. Support services and resources are signposted at the beginning of this guidance. If you find anything in this section distressing and would like to speak to someone or seek support, or if you or someone you know is struggling with suicide, self-harm or an eating disorder, please consider seeking help from the following resources: Samaritans Phone: 116 123 Shout Text: text ‘Shout’ to 85258 (free (free 24/7 helpline) support 24/7)

Childline Phone: 0800 1111 (for children and Beat the UK’s eating disorder charity. young people under 19) (24/7 helpline) Phone: 0808 801 0677 Campaign Against Living Miserably (CALM) (open from 3pm-8pm, Monday to Friday). Phone: 0800 58 58 58 NHS mental health services page by nation: (5pm to midnight every day) • NHS England PAPYRUS (for people under 35) Phone: 0800 068 41 41 (24/7 helpline) • NHS Scotland

SOS Silence of Suicide Phone: 0808 115 1505 • NHS Wales (8am to midnight weekdays and 4pm to • HSC Northern Ireland midnight weekends) Dial 999 when a crime is in progress or when someone is in immediate danger. This document is Ofcom’s Guidance on Content Harmful to Children for the purposes of section 53(1) of the Online Safety Act 2023 (the Act). The Act introduces a new concept of ‘content that is harmful to children’ which is relevant to the children’s risk assessment duties and children’s safety duties for services.

Overview¶

1.1 Section 53(1) of the Online Safety Act 2023 (the Act) requires Ofcom to provide guidance for providers of Part 3 services (i.e. user-to-user and search services) which gives examples of content, or kinds of content, that we consider to be, or consider not to be, primary priority content that is harmful to children (PPC) and priority content that is harmful to children (PC).

1.2 This Guidance on Content Harmful to Children is intended to support providers of Part 3 services that are likely to be accessed by children in making judgements about whether content on their service is content that is harmful to children as defined in the Act.

1.3 This guidance comprises nine sections, each of which provide examples of the content or kinds of content that Ofcom considers to be PPC and PC as defined in the Act. This structure broadly reflects the structure of the Children’s Register of Risks (Children’s Register): • Section 2: Guidance on pornographic content. • Section 3: Guidance on suicide content. • Section 4: Guidance on self-harm content. • Section 5: Guidance on eating disorder content. • Section 6: Guidance on abuse and hate content. • Section 7: Guidance on bullying content. • Section 8: Guidance on violent content. • Section 9: Guidance on harmful substances content. • Section 10: Guidance on dangerous stunts and challenges content.

1.4 Table 1.1 in this introductory section provides more detail on how these kinds of content that are harmful to children are defined in the Act.

1.5 In this section, we set out: • Legal framework: An overview of the legal framework of the Act that outlines Ofcom’s duties in producing guidance for service providers. This also outlines an explainer of service providers’ duties regarding content that is harmful to children. • Structure of this guidance: This includes an explanation of the structure of the guidance, including considerations for service providers, and the definitions relevant to each sub-section. • Freedom of expression: Information on the interaction of this guidance with freedom of expression.

Guidance in context¶

1.6 The Guidance on Content Harmful to Children interacts with other regulatory publications under the Act as follows: a) Children’s Risk Assessment Guidance for Service Providers (Children’s Risk Assessment Guidance): The guidance helps providers interpret the definitions of each kind of PPC and PC, facilitating compliance with the risk assessment duties, which requires assessing the level of risk for each kind of PPC and PC separately.

b) Children’s Register and Children’s Risk Profiles: As part of conducting their own risk assessments, services must consult the Children’s Risk Profiles, and we recommend services also consult the Children’s Register for a more in-depth understanding of the harms. Here, we compile relevant evidence to assess the presence and impact of each kind of harmful content, before identifying ‘risk factors’ (such as functionalities, user base characteristics or governance structures). The evidence presented in the Children’s Register and risk factors identified in the Children’s Risk Profiles inform our examples in the guidance. c) Protection of Children Codes: The Protection of Children Codes set out the user-to-user and search services codes of practice for providers of Part 3 services, describing measures recommended for compliance with specified duties. The harms guidance aids providers in assessing whether a piece of content constitutes PPC/PC and therefore whether Codes measures should be applied to prevent/protect children from encountering that content (see Protection of Children Codes). More generally, understanding the definitions of different types of PPC/PC will help providers meet their duties to have regard to freedom of expression when deciding on and implementing any safety measures and policies.

How the guidance will be updated¶

1.7 Due to the evolving nature of harmful content, this guidance will need to be updated as appropriate. We will monitor its effectiveness and proportionality at appropriate intervals, taking into account the evolving online context.

Legal framework¶

1.8 Section 236(1) of the Act defines content as anything communicated by means of an internet service, whether publicly or privately, including written material or messages, oral communications, photographs, videos, visual images, music and data of any description.

1.9 The Act defines content that is harmful to children as: a) Primary Priority Content (PPC); b) Priority Content (PC); or c) Other content which presents a material risk of significant harm to an appreciable number of children in the UK, which is sometimes referred to as ‘non-designated content that is harmful to children’ (NDC).1

1.10 NDC, as set out in (c), is a distinct category of content in the Act. It is not addressed in the Guidance on Content Harmful to Children but is addressed in the Children’s Register. In accordance with their children’s risk assessment duties, service providers are required to consider types of content that may be harmful to children beyond the designated harms specified by the Act.2 Providers should refer to the Introduction in Section 1 of the Children’s Register and the Children’s Risk Assessment Guidance for further detail on how they should consider NDC with regard to their children’s risk assessments duties.

1.11 The types of PPC and PC as defined by the Act that Ofcom is required to produce guidance on are set out in Table 1.1

1 Section 60(2) of the Act. 2 Sections 11 and 28 of the Act.

Table 1.1: Content harmful to children covered in our guidance as defined in the Act¶

Primary priority

Section

content that is harmful

Kind of harmful content, as defined in the Act

to children

Pornographic content,3 other than content which:

(a) Consists only of text, or

(b) Consists only of text accompanied by

Identifying content which consists only of text, (i)

(ii) Other identifying content which is not itself

Pornographic content

pornographic content,

A GIF which is not itself pornographic content, (iii)

An emoji or other symbol, or (iv)

(v) Any combination of content mentioned in sub-

paragraphs (i) to (iv).4

Content which encourages, promotes or provides instructions for

Suicide content

suicide.5

Content which encourages, promotes or provides instructions for an act

Self-injury content

of deliberate self-injury.6

Eating disorder

Content which encourages, promotes or provides instructions for an

content

eating disorder or behaviours associated with an eating disorder.7

Priority content that is

Kind of harmful content, as defined in the Act

harmful to children

Content which is abusive and which targets any of the following

characteristics:

(a) Race,

(b) Religion,

Content which is

(d) Sexual orientation,

abusive or incites

(e) Disability, or

hatred

(f) Gender reassignment.8

Content which incites hatred against people:

(a) Of a particular race, religion, sex or sexual orientation,

(b) Who have a disability, or

Content which encourages, promotes or provides instructions for an act

of serious violence against a person.10

Violent content

Content which:

(a) Depicts real or realistic serious violence against a person;

(b) Depicts the real or realistic serious injury of a person in graphic

detail.11

Content which:

(a) Depicts real or realistic serious violence against an animal;

(b) Depicts the real or realistic serious injury of an animal in graphic

detail;

or the serious injury of a fictional creature in graphic detail.12

Bullying content.13

Content may, in particular, be bullying content if it is content targeted

Bullying content

against a person which –

(a) Conveys a serious threat;

(b) Is humiliating or degrading;

Content which encourages, promotes or provides instructions for a

Dangerous stunts and

challenge or stunt highly likely to result in serious injury to the person

challenges content

who does it or to someone else.15

Content which encourages a person to ingest, inject, inhale or in any

Harmful substances

other way self-administer:

content

(a) A physically harmful substance;

(b) A substance in such a quantity as to be physically harmful.16

3 Section 61(2) of the Act. 4 Section 61(6) of the Act.

5 Section 61(3) of the Act. 6 Section 61(4) of the Act. 7 Section 61(5) of the Act. 8 Section 62(2) of the Act. 9 Section 62(3) of the Act. 10 Section 62(4) of the Act. 11 Section 62(6) of the Act.

1.12 This document intends to provide services with helpful guidance on making judgements on PPC and PC, rather than an attempt to anticipate every circumstance which may arise during content moderation. This guidance also includes further contextual information for services to assist when assessing particular content types.

1.13 Section 192 of the Act sets out the approach to be taken where either a system or process operated or used by a service to comply with the Act, or a risk assessment, involves a judgement by the service about whether content is harmful to children. In summary, such judgements are to be made on the basis of all relevant information reasonably available to the provider.

Service providers’ duties regarding content that is harmful to children¶

1.14 The Act creates new legal duties for providers of Part 3 services likely to be accessed by children in relation to content that is harmful to them. This guidance is relevant to such services in the carrying out of their duties.

1.15 These duties include, amongst others, duties on both user-to-user and search services to carry out suitable and sufficient children’s risk assessments and to take proportionate measures relating to the design and operation of the service to effectively mitigate and manage the risks of harm identified in those risk assessments. They also include a duty on user-to-user services likely to be accessed by children to operate a service using proportionate systems and processes designed to prevent children of any age from encountering PPC that is harmful to children and to protect children in age groups judged at

12 Section 62(7) of the Act. 13 Section 62(5) of the Act 14 Section 62(12) of the Act. 15 Section 62(8) of the Act. 16 Section 62(9) of the Act.

risk of harm from PC and NDC from encountering it. Search services likely to be accessed by children are subject to a duty to use proportionate systems and processes designed to minimise the risk of children of any age from encountering PPC and minimise the risk of children in age groups judged to be at risk of harm from PC and NDC from encountering it. The guidance aims to help service providers make judgements that may be required to comply with the requirement to conduct risk assessments regarding content harmful to children under the children’s risk assessment duties (see Children’s Risk Assessment Guidance), or to comply with their children’s safety duties which they may do by implementing measures set out in Ofcom’s Protection of Children Codes (see Protection of Children Codes).

1.16 When services conduct risk assessments and implement measures to comply with the relevant duties, they are likely to be dealing with categories of content, as opposed to making an assessment on an individual piece of content. Providers should anticipate that some of the content could be harmful to children. However, for compliance purposes – for example, in content moderation – service providers will need to make decisions about specific pieces of content. This guidance will help services design their content moderation policies to appropriately deal with broad categories of harmful content.

1.17 Services are likely to be making content moderation judgements to comply with the laws of every country in which they operate, and different countries have different laws. This guidance is not intended to override or supersede existing moderation practices, where these practices already meet the duties set out in the Act. Rather, this guidance sets out how service providers need to consider content that is harmful to children in order to comply with the requirements of the children’s safety duties. Please see the Protection of Children Codes for more detail.

Structure of this guidance¶

1.18 There is one section for each of the nine categories of content.

1.19 Each section includes: • Relevant illegal content. • Key terms. • Additional context for service providers. • Examples.

1.20 The following sub-section outlines the structure in more detail.

Relevant illegal content¶

1.21 The ‘Relevant illegal content’ sub-section explains how some of the kinds of content harmful to children are closely linked to the offences outlined within the Illegal Content Judgements Guidance (ICJG). This Guidance on Content Harmful to Children does not, and cannot, set out to definitively demarcate between illegal content and content that is harmful to children (within the meaning of the Act). It is also not the purpose of this guidance to reiterate material that is captured within the ICJG. Content that we consider to be illegal content, as outlined in the ICJG, is not included in our examples in this guidance.

1.22 We recognise that it can be difficult to make distinctions between illegal content and content that is harmful to children. Service providers should be mindful that different safety duties apply to content that is illegal, and that these duties apply to all users. Where services are considering content that could fall in scope of the definitions of content that is harmful to children, but have reasonable grounds to infer that content is illegal, they should apply the illegal content safety duties as these will provide the greatest level of protection to all users, including children, in accordance with the Act.

1.23 Within each section, where categories of PPC or PC are associated with illegal content, we set out the circumstances in which content may be illegal, and signpost to the ICJG section service providers should refer to, to understand the criteria for illegality in full.

Key terms¶

1.24 The ‘Key terms’ sub-section helps service providers identify certain overarching terms used to define specific types of content. For some harmful content, the definitions in the Act are dependent on descriptive terms such as ‘promotes’. We have therefore provided examples for different kinds of harmful content, to demonstrate how we interpret these terms.

Additional context for service providers¶

1.25 This sub-section provides information to service providers about how contextual factors may be relevant to determining whether or not content falls within definitions of harmful content as it may not always be clear whether content does or does not meet the definitions of PPC or PC.

1.26 We encourage service providers to take care in instances in which it may not be immediately clear whether content meets the definition of PPC or PC and give due regard to additional context available: such context might be important in determining whether or not content should be considered harmful to children. In particular, this assessment may depend on additional context relating to the nature of the content, or how it is presented on the service. This could include any code words, substitute terms and phrases, hashtags, emojis17, sounds and comments that might relate to it, as well as other relevant information reasonably available to the provider, such as information about the user who has shared the content or how they shared it (e.g., the group, channel or community on which it has been shared).

1.27 Within each section we set out the key contextual considerations that services should look out for when assessing different kinds of content that is harmful to children. For example, while an underlying image or video posted to the service may not meet the definition of content that is harmful to children when considered in isolation, we explain that in some cases the hashtags or trending songs that accompany it may indicate that it is, in fact, a kind of content that is harmful to children.

1.28 It may also not always be immediately clear whether ‘recovery content’ – which we refer to in our suicide, self-harm18 and eating disorder sections – will meet the definition of PPC.

17 However, pornographic content that is harmful to children specifically excludes content that consists only of text (i.e. written material) accompanied by emojis or other symbols. See sections 236(1), 61(6)(a) and 61(6)(b)(i) – (v) of the Act for the full definition of pornographic content and relevant exclusions. 18 Throughout this section and the rest of the guidance, we mostly use the established and widely understood term ‘self-harm’ to refer to “an act of deliberate self-injury” as specified in the Act.

Recovery content presents a particular challenge in this regard because while some of this content may serve a valid purpose for self-expression and aiding recovery, it may still meet the definition of PPC in the Act. For example, harmful content might be shared by those with little medical or professional expertise in suicide, self-harm and eating disorders, and who are themselves experiencing poor mental health. In addition, posts that are not themselves harmful may attract comments that encourage others to engage in suicide, self-harm or behaviours associated with eating disorders (e.g., because they provide instructions for or explain methods tried by other users). Some recovery content may, therefore, meet the definition of PPC even where the person who shares it does not intend to be harmful.

1.29 Service providers should recognise that there is a balance to be struck in distinguishing the circumstances in which content may or may not meet the definitions under the Act. Service providers should consider the contextual considerations set out within each section of this guidance, as these may assist with assessments of different kinds of content which may be borderline harmful.

1.30 In some instances, there may also be overlaps between different categories of PPC and/or PC that providers will need to consider. For example, suicide content may overlap with other categories of content, such as dangerous stunts and challenges content or harmful substances content, where online challenges dare children to carry out lethal self-harm or suicide methods (see Sections 9 and 10).

Examples¶

1.31 The remaining sub-sections contain non-exhaustive lists of examples of content, or kinds of content, that we consider to be, or consider not to be, PPC or PC. For each kind of harmful content, we have opted to include tables with ‘content descriptions’ or ‘kinds of content’ alongside more specific examples to illustrate harmful content.

Freedom of expression¶

1.32 Ofcom must carry out its functions compatibly with fundamental rights in accordance with its duties under the Human Rights Act 1998, including the right to freedom of expression under Article 10 of the European Convention on Human Rights.19 Any limitation on the right to freedom of expression must be prescribed by law, pursue a legitimate aim and be necessary in a democratic society. In order to be ‘necessary’, the restriction must correspond to a pressing social need, and it must be proportionate to the legitimate aim pursued. Both the definition of content that is harmful to children and the requirement for Ofcom to prepare this guidance are set out in the Act and pursue the legitimate aim of protecting children from harm online. We have had careful regard to the right to freedom of expression in producing this guidance.

19 Article 10(1) states that the right to freedom of expression includes the freedom to hold opinions and to receive and impart information and ideas without interference by public authority. Article 10(2) states that this right may be restricted in the interests of national security, territorial integrity or public safety, for the prevention of disorder or crime, for the protection of health or morals, for the protection of the reputation or rights of others, for preventing the disclosure of information received in confidence, or for maintaining the authority and impartiality of the judiciary.

1.33 It is open to services as a commercial matter (and in the exercise of their own rights to freedom of expression) to allow or prohibit content that is harmful to children, so long as they abide by the Act. For example, some services likely to be accessed by children may choose to allow adult users to share particular types of PPC on a service. It is open to them to do this. This guidance is intended to help service providers identify particular content types to ensure that, regardless of whether it may or may not be prohibited on a service, they take appropriate measures to ensure that the relevant children’s safety duties are met, and child users are protected or prevented from encountering such content. Service providers will need to ensure in doing so that they also comply with data protection law.20

1.34 Service providers should be mindful that sections 22 and 33 of the Act set out that when services are deciding on implementing safety measures and policies, they should have particular regard to users’ rights of freedom of expression. By including examples within this guidance that do attract protection under the Human Rights Act 1998, we are not suggesting that these forms of content should be taken down or disallowed from the service. Rather, such content should be treated in accordance with the relevant duties.

1.35 We are mindful that certain forms of speech attract a high degree of protection under the right to freedom of expression. We acknowledge that there needs to be a strong justification in the public interest for restricting certain forms of speech under Article 10, including for children. For example, we are mindful that journalistic content, artistic content and educational content all warrant careful consideration in particular.21 As such, we have set out more detail on these content types below and throughout our guidance, we have sought to exemplify the circumstances in which this content may, or may not, meet the definition of PPC and PC.

Journalistic content and content of democratic importance¶

1.36 Journalistic content and content of democratic importance, as defined in the Act,22 could potentially meet the definition of one of more kinds of PPC or PC.

1.37 The Act sets out that ‘journalistic content’ includes ‘news publisher content’.23 While journalistic content is subject to the Part 3 duties relating to protection of children, news publisher content is exempt from these duties (as well as the duties that apply to illegal content).24 We therefore focus on the broader category of ‘journalistic content’ in this guidance.

1.38 When determining whether content is news publisher content or journalistic content and is therefore subject to the Part 3 duties, service providers should refer to the relevant statutory definitions25 and, in particular, consider the source and purpose of content.

1.39 If the protection of children duties do not apply (e.g., where content is produced by a recognised news publisher), services are not expected to take any further action.

20 Information Commissioner’s Office. A guide to the data protection principles. [accessed 26 March 2025]. 21 House of Lords, 2004. Judgments – Campbell (Appellant) v. MGN Limited (Respondents), Session 2003-04. [accessed 28 March 2025]

22 Sections 17(7) and 20(10) of the Act. 23 Section 19(10(a)) of the Act. 24 Sections 55-56 and 57(2)(b) and (c) of the Act. News publisher content is excluded from the definition of regulated user-generated content and search content in the Act, and is therefore not subject to the safety duties about protecting children and illegal content. 25 Sections 56 and 19(10).

1.40 If the protection of children duties apply (e.g., where content is produced by freelance or citizen journalists), service providers should consider whether the content meets the definition of content that is harmful to children as explained in the sections to follow. Alternatively, they may apply the provisions of their terms or publicly available statements, where such provisions are broad enough to cover such content. Further, services should consider their sections 22 and 33 duties as set out above.

Artistic and educational content¶

1.41 We are particularly mindful of the rights of children to have access to artistic and educational content as these are important both for their social and educational development. Nonetheless, some artistic and educational content may meet the definitions of PPC and PC under the Act, because they are likely to be harmful to children. We have therefore provided examples in each section where this may be the case, as well as providing examples of where this is not likely to be the case.

Section 2 Guidance on pornographic content¶

Warning: this section contains references to content that may be upsetting or distressing, including discussions of sexual violence. Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be pornographic content. Pornographic content means content of such a nature that it is reasonable to assume that it was produced solely or principally for the purpose of sexual arousal.

Considerations for service providers¶

2.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).26 Pornographic content that is published or displayed on an online service by the provider of the service is subject to different duties under Part 5 of the Act, and Ofcom has published guidance for providers of this content. The scope of the Part 3 and Part 5 duties differs based on how the content is shared, published or displayed, rather than differences in the definition of pornographic content itself. Therefore, the examples of content or kinds of content that fit into the category of ‘pornographic content’ will be the same for Part 3 and Part 5.

2.2 In this section, we outline examples of content that Ofcom considers to be, or not to be, pornographic content, a form of primary priority content that is harmful to children (PPC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

Relevant illegal content¶

2.3 We have provided comprehensive guidance on how services should make illegal content judgements in relation to a range of image-based sexual offences in the Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online, and services should refer to this to understand the criteria for illegality in full. All illegal sexual content as described in paragraphs 2.4-2.6 falls outside the scope of this guidance.

2.4 Broadly, content may be illegal where there are reasonable grounds to infer that it: a) Includes material which meets the definition of ‘Extreme pornography’ (see section on the ‘Extreme pornography’ offence in the ICJG). Service providers should note that simulated acts, including acts depicted in synthetic images such as ‘deepfakes’, can fall within the definition of ‘extreme pornography’, provided they are depicted in an explicit and realistic way.

26 Section 53 of the Act.

b) Involves the non-consensual disclosure of, or threats to disclose, intimate images (see section on the ‘Intimate image abuse’ offence in the ICJG). c) Involves a user intentionally sending a photograph or film of genitals for the purposes of causing alarm, distress or humiliation or for the purpose of obtaining sexual gratification (see section on the ‘Cyberflashing’ offence in the ICJG).

2.5 In addition, images or videos which show a child engaged in sexual activity, or appearing to be engaged in sexual activity, constitute child sexual abuse material (CSAM), which is illegal content (see offences related to indecent and prohibited images of a child in the ICJG section on CSAM).

2.6 Pornographic content may also be used in the commission of criminal offences, including as part of the grooming process, to coerce a child into performing sexual acts, or causing a child to watch a sexual act for the purposes of sexual gratification. Guidance for identifying these offences online can also be found in the ICJG.

Key terms¶

Table 2.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Sexual activity

Depictions of consensual sexual acts such as masturbation, oral sex,

penetration or ejaculation.

Pornographic content

Pornographic content means content of such a nature that it is reasonable to assume that it was produced solely or principally for the purpose of sexual arousal.27 For the purposes of the Act, pornographic content that is harmful to children specifically excludes content that (a) consists only of text (i.e., written material);28 or (b) consists only of text and is accompanied by identifying content (that may be text or another kind of content which is not itself pornographic), non-pornographic GIFs, emojis or other symbols, or any combination of these.29

Sexually suggestive or sexualised content

Any material that implies or hints at sexual themes and/or intention, to evoke sexual thoughts or reactions, but that is not explicit and has not been produced solely or principally for the purpose of sexual arousal, and therefore does not meet the definition of pornographic content (e.g., flirtatious or provocative material, double entendre and innuendo).

Additional context for service providers¶

2.7 To assess whether content is pornographic is to understand whether it is reasonable to assume that it was produced ‘solely or principally’ for the purpose of sexual arousal. Whether content can be assumed to have been produced either solely or principally for the

27 Section 236(1) of the Act. 28 Section 61(6)(a) of the Act. 29 Section 61(6)(b)(i) to (v) of the Act.

purpose of sexual arousal is dependent on the nature of the content itself, having taken contextual factors into account, rather than the intent of the uploading user or any viewer of it.

2.8 The presence and combination of particular elements within content will make it more or less likely for the content to be deemed as having the principal or sole purpose of sexual arousal; these elements include:

• More likely > Presence of nudity, especially full frontal or detailed breast or genital nudity. > References to or depictions of fetish material. > Depictions of real sexual activity. > Very strong references to sexual behaviour:  use of language associated with sexual activity and pornography, such as ‘milf’, ‘horny’;  simulations or suggestions of sexual activity, such as simulations of oral sex or sexually suggestive moaning;  depictions of symbols or objects associated with sex, such as sex toys. • Less likely > A clear alternative purpose for the content, such as an artistic, dramatic, educational or other non-sexual purpose. > Semi-nudity, or the individuals in the content are clothed and not in underwear or fetish clothing.

2.9 Assessments of content should be holistic. For example, content may depict individuals who are fully clothed, but if they are simulating sexual activity and using pornographic language, without indication of a clear alternative purpose for the content, this should be deemed as pornographic content. However, if content depicts an individual using pornographic language but they are fully clothed and not carrying out or simulating sexual activity this would likely not be judged to be pornographic.

2.10 For content to be considered pornographic rather than sexually suggestive, the references to sexual behaviour must be strong, such that it is clear and obvious that the content is intended for the purpose of sexual arousal. This purpose can often be identified through a combination of factors, such as nudity or partial nudity or references to fetish material. Similarly, references to common fetishes may not be sufficient for content to be clearly intended for sexual arousal. For example, an image of an individual’s feet without any other indicators that the purpose of the image is for sexual arousal could feasibly be for an alternative purpose and should not be considered pornographic content. However, when combined with additional factors, such as the presence of nudity or partial nudity, suggestions or simulations of sexual activity or use of pornographic or sexual language could contribute to the overall judgement that content is pornographic.

2.11 Content that would be classified as pornographic by the British Board of Film Classification (BBFC) or Ofcom when appearing on other forms of regulated media, should be considered pornographic when shared by users. This might include:

• A clip from broadcast television depicting graphic simulations of sexual activity clearly intended to elicit sexual arousal; a clip from a film rated R1830 or sex work at 1831 by the BBFC. • A clip that would be deemed adult sex material by Ofcom when appearing on broadcast television or an on-demand programme services.32 33 • A photograph that would be deemed sexual content unsuitable for under 18s when accessed via a mobile network.34

2.12 Pornographic content may include synthetic image, video and audio content, for example: • Artificial intelligence (AI)-generated images and videos, such as images of consenting35 real people, that are consensually shared, or ‘hyper realistic’ generated images of people who are not real. • Animated images or videos, such as from video games, hentai or cartoons, depicting detailed nudity and/or sexual activity. • Paintings, drawings or other non-photographic images.

Examples¶

Examples or kinds of content that Ofcom considers to be pornographic content that is harmful to children¶

2.13 Table 2.2 describes the kinds of pornographic content that Ofcom considers to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

2.14 We have not addressed the level or impact of harm within Table 2.2. For further information on the risks and impact of harm, see Section 2 of the Children’s Register of Risks.

2.15 Services should keep in mind that the kinds of content must be solely or principally for the purposes of sexual arousal, in accordance with the definition of pornographic content in the Act. If the content is not for the purposes of sexual arousal, and is instead for a medical, educational, artistic or other non-sexual purpose, it will not meet the definition of pornographic content in the Act.

2.16 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

30 BBFC, n.d. R18 rating. [accessed 16 April 2024]. 31 BBFC, n.d. 18 rating. [accessed 16 April 2024].

32 Ofcom, 2021. The Ofcom Broadcasting Code. Section one: Protecting under-eighteens. [accessed 16 April 2024]. 33 Ofcom, 2021. On-demand programme services (“ODPS”) guidance. [accessed 16 April 2024]. 34 BBFC, n.d. The BBFC Mobile Content Framework. [accessed 16 April 2024]. 35 Non-consenting ‘deepfakes’ may be considered non-consensual intimate image (NCII) abuse, a criminal offence contrary to section 66B of the Sex Offences Act 2003 as amended by section 188 of the Online Safety Act 2023 when they are shared.

Table 2.2: Descriptions and examples of pornographic content that is harmful to children¶

Content description

Content examples

Explicit depictions of sexual activity

• Depictions of sexual acts such as masturbation, oral sex, penetration or ejaculation, including where these acts are performed with or on a sex toy.

Content depicting full frontal nudity or depicting genitals, breasts or buttocks to elicit sexual arousal

• Any content focusing on the breasts (with or without

exposed nipple, depending on the context), genitals

or buttocks or depicting full frontal nudity.

• Videos showing an individual with exposed or partially

exposed genitals, breasts or buttocks where they are

using language associated with sexual activity or

pornography (e.g., ‘milf’ or ‘horny’), or strongly

suggesting sexual activity (e.g., moaning or simulating

masturbation or oral sex).

• An autonomous sensory meridian response (ASMR)

video where sexually suggestive sounds are combined

with full or partial nudity.

Fetish material

• Depictions of an individual with their breasts, buttocks or genitals exposed depicting bondage, discipline, sadism and/or masochism (BDSM) or any other fetish depictions. • Depictions of an individual being whipped or restrained containing language associated with sexual activity or pornography (e.g., ‘milf’ or ‘horny’),36 or other strong suggestions of sexual activity (e.g., moaning or simulating sexual activity).

Explicit audio material

• Audio material includes graphic descriptions of sexual activity and/or sexual soundscapes, (e.g., moaning combined with rhythmic bodily contact sounds) that are sexual in nature and are intended solely or principally for sexual arousal.

Examples or kinds of content that Ofcom considers not to be pornographic content that is harmful to children¶

2.17 Table 2.3 describes the kinds of pornographic content that Ofcom does not consider to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

2.18 The kinds of content in Table 2.3 are either (a) out of scope (text-based content that is expressly excluded in the Act), or (b) deemed to be content for which it is reasonable to assume the sole or principal purpose is not sexual arousal.

36 We are aware that use of language, particularly in online spaces, evolves and develops over time. The examples provided here are contemporary at the time this guidance has been written.

2.19 Services should be mindful that the examples in the table must not be for the sole or principal purposes of arousal. If the primary purpose is artistic, for example, content will not meet the definition under the Act.

2.20 Although Ofcom does not consider the examples in the following table to meet the definition of PPC, the content may still be inappropriate for children in certain age groups.

Table 2.3: Descriptions and examples of content that are not pornographic¶

Content description

Content examples

Text-based content of a sexual nature: this includes text-based content of a sexual nature accompanied by GIFs (that are not sexual in nature), emojis or symbols

• Written erotica or fan fiction. • A non-sexualised emoji or GIF alongside text-based content of a sexual nature. • Audiobook material/erotica in which character development, narrative structure and/or unabridged adaptation of the written word means that the sole or principal purpose of the work is not for sexual arousal.

Glamour content whose primary

• A video of someone playing a video game in swimwear. • Sexualised or suggestive content of a kind that might be expected to feature in an advertisement.

purpose falls short of sexual

arousal, but may have suggestive

intentions

Content depicting partial or full nudity in a non-sexual context

• Content depicting breastfeeding. • Content depicting nudity for a comedic or satirical purpose. • Journalistic content including images of non-sexual nudity.

Educational material which includes imagery of, or discussion about, anatomy, nudity or sexual activity

• Anatomical diagrams. • Content intended for the sex education of children. • Depictions of penetration in a medical context, e.g., a demonstration of an intimate examination. • Academic material, e.g., historical, sociological or anthropological content, presented and appropriately contextualised as one might expect to find in an encyclopaedia or academic journal.

Dramatic or comedic content where nudity or sexual activity is secondary to the dramatic or comedic purpose of the content

• Depictions of nudity or sexual activity that are not

shown in graphic detail37 and where a comedic or

dramatic purpose is clear. This may include clipped

footage of a sex scene from film or television, where

the sexual activity is not shown in graphic detail and

the dramatic purpose is evident and retained in the

clip, e.g., through a combination of factors such as the

inclusion of a storyline, character development and/or

narrative device.

Content description

Content examples

Artwork featuring nudity or sexual activity where the primary purpose is artistic

• Content depicting sexually suggestive dancing or acrobatics (e.g., dancing in music videos, pole dancing or aerial acrobatics). • An image of a painting, sculpture, photograph or other artwork containing nudity.

Content associated with common or known fetishes, but that is not graphic or overtly sexual in nature

• Videos or images of feet with no additional context to

suggest a sexual purpose.

• Depictions of individuals in commonly sexualised fancy

dress (e.g., fur suits, nurses, maids, butlers, police

officers).

Marketing material and depictions used for the promotion and sale of sex toys and sexual wellness products

• Depictions that focus on a sex toy or sexual wellness product ‘as is’ rather than ‘in use’, and that do not show full frontal nudity or genitalia. • Descriptions of the efficacy and features of sex toys or sexual wellness products, that are designed for the promotion, sale and/or consumption of the product, rather than sexual arousal.

This guidance relates to duties for user-to-user and search services under Part 3 of the

Online Safety Act 2023 (the Act).39 In this section, we outline examples of content that

Ofcom considers to be, or not be, suicide content, a form of primary priority content that is

harmful to children (PPC) under the Act. For the purpose of this guidance, all mentions of

image, video or audio content included in the tables refer to both real and synthetic images,

videos or audio content.

37 In line with the BBFC 15 classification. Source: BBFC, n.d. 15 rating. [accessed 16 April 2024].

Section 3 Guidance on suicide content¶

Warning: this section contains references to content that may be upsetting or distressing, including detailed discussion and descriptions of suicide and examples of suicide content. Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be content which encourages, promotes or provides instructions for suicide.38 This is referred to as suicide content.

Considerations for service providers¶

3.1

3.2 People who are posting or interacting with suicide content may be vulnerable or in difficult circumstances. We are conscious of the stigma which may be associated with suicidal ideation and that this may present a barrier to individuals seeking help. It is important for children to have the opportunity to access and share content about their mental health. We have been mindful of this when developing this guidance and would strongly recommend service providers look at Section 3 of the Children’s Register of Risks (Children’s Register) in conjunction with this guidance to understand in more detail the sensitivities around this content.

3.3 Ofcom does not expect services to restrict children’s access to genuine safety-promoting, non-harmful information and resources relating to suicide.

Relevant illegal content¶

3.4 Suicide content may be illegal in some circumstances. Broadly, content should be considered as illegal suicide content where there are reasonable grounds to infer that it amounts to an offence of intentionally encouraging or assisting the suicide (or attempted suicide) of another person.40

3.5 We have provided comprehensive guidance on how services should make judgements in relation to illegal suicide content in Section 13 (Encouraging or assisting suicide) of the Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online. Services should refer to that guidance to understand the criteria for illegality in full.

38 Section 61(3) of the Online Safety Act 2023. 39 Section 53 of the Act. 40 Section 2 of the Suicide Act 1961; Section 13 of the Criminal Justice Act (Northern Ireland) 1966.

Key terms¶

Table 3.1: Terms frequently used in this section of the guidance, or terms listed in the Act¶

Term

Definition

Encourages

Content which could persuade others to contemplate suicide and/or

make others more likely to attempt or consider suicide as a course of

action.

This includes content which glamourises, glorifies, romanticises or

normalises suicide.

Unlike illegal suicide content (please refer to the ICJG), suicide content

that is harmful to children does not need to intentionally encourage

suicide.

Promotes

Content which publicises, supports or recommends suicide as an act (this does not need to be intentional or explicit). This includes content which glamourises, glorifies, romanticises or normalises suicide.

Provides instructions

Content which describes a method for suicide in sufficient detail (such

for

as setting out steps or diagrams) that can be emulated.

Suicide

The act of an individual intentionally ending their own life.

Additional context for service providers¶

3.6 In some instances, there may be overlaps between suicide content and other categories of content such as self-harm and eating disorder content (see Sections 4 and 5), bullying content (see Section 7), harmful substances content (see Section 9) and dangerous stunts and challenges content (see Section 10). For example, this could include online challenges which dare children to carry out lethal self-harm or suicide methods.

3.7 Service providers should also be mindful where content relating to suicide provides safety-promoting resources and should ensure that children are able to access resources from genuine support services. Children should be able to access responsible, health-promoting information about, and support for, suicidal ideation.

3.8 Service providers should also carefully consider recovery content when making judgements. Recovery content can take different forms for different users and can play an important role in providing hope and challenging stigmas. However, some content which is intended to be or is presented as ‘recovery content’ will nevertheless be harmful to children. Service providers should be aware that: a) Some children, including those in crisis or who are particularly vulnerable, are more likely to be encouraged or instructed by some recovery content regardless of the intention behind the post. This could include detailed documentations of recovery journeys which could provide instructions for suicide. b) Harmful comments may be posted on genuine recovery content, for example, when users ‘take over’ another user’s post and flood it with harmful comments about suicide.

c) Harmful content about suicide may be posted in genuine recovery forums, groups or communities.

3.9 In addition, service providers should be aware that the presentation of suicide content develops quickly and evolves rapidly, which can enable such content to evade content moderation. This can include the use of code words and hashtags employed by users to signal and search for suicide and self-harm content. Often, these terms can be specific to a particular service. Some content may also adapt commonly used terms to obscure the content (e.g., intentional misspelling/leet,41 algospeak,42 use of acronyms, the adoption of coded words originating in popular culture). Service providers should therefore keep in mind that while a post might not explicitly suggest suicide, it could be accompanied by more suggestive comments, captions or hashtags which indicate that the content does in fact meet the definition of suicide content.

3.10 Service providers should take care when assessing artistic representations of suicide and self-harm, for example, music, poetry, artwork or fictional works, drawings, stories, paintings and other art. These may romanticise or glamourise suicide in a way that is harmful to children. We are mindful of the need to consider freedom of expression. However, where content meets the definition of suicide content under the Act, then it should be treated as suicide content, regardless of the format. Service providers should therefore focus on whether the content meets the definition under the Act.

Examples¶

Examples or kinds of content that Ofcom considers to be suicide content that is harmful to children¶

3.11 Table 3.2 describes the kinds of suicide content that Ofcom considers to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

3.12 We have not addressed the level or impact of harm within the table. For further information on the risks and impact of suicide content, see Section 3 of the Children’s Register.

3.13 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in Table 3.2.

3.14 Service providers should keep in mind that, in accordance with the Act, the kinds or types of content listed must encourage, promote or provide instructions for suicide, but do not need to do so intentionally.

41 Leet speak or ‘l337 speak’ refers to an informal online language where numbers or special characters are used to replace vowels or consonants. 42 Algorithm speak or ‘algospeak’ refers to coded language used online to circumvent content moderation methods. This could also include the use of emojis.

Table 3.2: Descriptions and examples of content which encourages, promotes or provides instructions for suicide

Content descriptions

Content examples

Content describing methods of suicide

• A post describing methods of suicide as: pain-free,

quick, instant, easy or common.

• A post describing an unusual, emerging or novel

method of suicide.

• A post describing particular methods as not leaving a

lasting effect on the body should a suicide attempt not

result in death, or as not being traceable.

• A video recommending specific places where

individuals have taken their own life.

Content containing detailed instructions for methods of suicide

• A post containing descriptions or depictions of detailed steps to carry out a particular method for suicide, such as a step-by-step guide. • A post describing lethal substances and other details such as dosages or how to obtain them to use for suicide. • Descriptions of an individual’s personal account of how they could or plan to (attempt to) take their own life.

Content which represents suicide in a way which romanticises, glamourises, glorifies or normalises it

• Content depicting common accessories for suicide,

accompanied by promoting text, sounds or hashtags.

• Content which portrays suicide as a positive way to

end distress.

• Content which describes suicide as something that is

inevitable.

• Content which describes suicide as a positive response

to problems.

Content, such as internet challenges or dares, instructing suicide

• Online challenges that dare or otherwise encourage users to engage in lethal self-harm or suicide methods. • Discussions of suicide pacts in comments or discussion forums where users encourage each other to take their own lives.

Content featuring real-life suicides or suicide attempts

• Content that shows any individual actively preparing

to take their own life, e.g., livestreaming or pre-

recorded. This could include them discussing their

plans in detail or preparing accessories, or tools for

suicide.

• Images of individuals where it is clear the individual

has died by suicide, e.g., the method is depicted.

Content descriptions

Content examples

Brief descriptions or mentions of

• Content which includes limited details such as ‘my uncle took his own life’.

suicide where the description is

limited, the method is well known,

and the content does not otherwise

encourage or promote

Content which promotes recovery and awareness of suicide in a healthy, non-harmful way and/or which reduces stigma

• Discussions of healthy coping mechanisms. • Descriptions of signs that a person may be experiencing suicidal ideation and how to assist them. • Explanations of how to talk about suicide in a non-stigmatising and safe way. • Information about how and where to seek further help. • Facts and statistics about suicide (but not methods). • Individuals discussing their own recovery, or seeking support.

Content which discusses an individual’s lived experience, which does not discuss detail or methods

• Content which discusses someone’s lived experience

of suicidal thoughts and feelings without detail or

commentary on the methods, and which does not

encourage or promote suicide.

Safety-promoting resources which provide healthy coping mechanisms or signpost to services for recovery

• Content from recognised suicide prevention bodies, such as charity resources and sites. • NHS website, National Institute for Health and Care Excellence (NICE) clinical guidelines and other recognised medical resources. • Government suicide prevention material. • Resources which encourage individuals to seek help or speak to someone, or refer individuals to places that they can seek help.

Academic or educational articles

• Articles related to suicide rates, or suicide prevention

methods.

Content descriptions

Content examples

Content that discusses belief and

• A post expressing a wish or belief that an individual who has died by suicide is in a better place.

hope in, e.g., afterlife, heaven or

eternal life in a religious or spiritual

context

Political or policy discussions

• Content debating or discussing assisted dying.

Examples or kinds of content that Ofcom considers not to be suicide content that is harmful to children.¶

3.15 The following table describes the kinds of suicide content that Ofcom does not consider to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

3.16 Although Ofcom does not consider the examples in the following table to meet the definition of PPC, the content may still be inappropriate for children in certain age groups.

3.17 Where content promotes recovery or awareness of suicide and does not contain any of the content listed in Table 3.2, this may not be considered PPC.

Table 3.3: Descriptions and examples of content which is not content that encourages, promotes or provides instructions for suicide

Section 4 Guidance on self-harm content¶

Warning: this section contains references to content that may be upsetting or distressing, including detailed discussion and descriptions of self-harm and examples of harmful content. Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be content which encourages, promotes or provides instructions for an act of deliberate self-injury.43 This is referred to as self-harm content.

Considerations for service providers¶

4.1 This guidance is required in relation to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).44 Throughout this section and the rest of this guidance we mostly use the established and widely understood term ‘self-harm’ to refer to “an act of deliberate self-injury” as specified in the Act. In this section, we outline examples of content that Ofcom considers to be, or not be, self-harm content, a form of primary priority content that is harmful to children (PPC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

4.2 People who are posting or interacting with self-harm content may be in vulnerable or difficult circumstances. We are conscious of the stigma which may be associated with self-harm and that this may present a barrier to individuals seeking help. It is important for children to access and share content about their mental health and we have been mindful of this when developing this guidance. We also expect services to review Section 3 of the Children’s Register of Risks (Children’s Register) in conjunction with this guidance to better understand the sensitivities around this content.

4.3 Ofcom does not expect services to restrict children’s access to genuine safety-promoting, non-harmful information and resources relating to self-harm.

Relevant illegal content¶

4.4 Self-harm content may be illegal in some circumstances. Content should be considered illegal where there are reasonable grounds to infer that it amounts to an offence of intentionally encouraging or assisting an act of serious self-harm (i.e., self-harm which would result in grievous bodily harm or severe injury, including psychiatric harm).

4.5 We have provided comprehensive guidance on how services should make illegal content judgements in relation to serious self-harm content in the ‘Encouraging or assisting serious self-harm’ sub-section of the Illegal Content Judgements Guidance (ICJG) published as part

43 Section 61(4) of the Online Safety Act 2023. 44 Section 53 of the Act.

of our December 2024 Statement on Protecting People from Illegal Harms Online. Service providers should refer to this to understand the criteria for illegality in full.

Key terms¶

Table 4.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Self-injury/self-harm

When someone hurts themselves as a way of dealing with very difficult

feelings, painful memories or overwhelming situations and

experiences.

Encourages

Content which could persuade others to contemplate self-harm and/or make others more likely to attempt or consider self-harm as a course of action. Unlike illegal self-harm content (please refer to the ICJG), self-harm content that is harmful to children does not need to intentionally encourage self-harm.

Promotes

Content which publicises, supports or recommends self-harm as an act

(this does not need to be deliberate or explicit).

This includes content which glamourises, glorifies, romanticises or

normalises self-harm.

Provides instructions for

Describes a method for self-harm in sufficient detail (such as setting out tools, steps or diagrams) that can be emulated. As with ‘encourages’, any description of a method of self-harm does not need to intentionally provide instructions for self-harm to be harmful to children. Instructions may be minimal and still able to be emulated by children.

Additional context for service providers¶

4.6 In some instances, there may be overlap between self-harm content and other categories of content such as suicide and eating disorder content (see Sections 3 and 5), bullying content (Section 7), harmful substances content (Section 9) and dangerous stunts and challenges content (Section 10). For example, this could include online challenges which dare children to carry out self-harm or suicide methods.

4.7 Service providers should also be mindful where content relating to self-harm provides safety-promoting resources and should ensure that children are able to access resources from genuine support services. Children should be able to access responsible, safety-promoting information about, and support for, self-harm ideation.

4.8 Service providers should also carefully consider recovery content when making judgements. Recovery content can take different forms for different users, and can play an important role in providing hope and challenging stigmas. However, some content which is intended to be or is presented as recovery content will nevertheless be harmful to children. Service providers should be aware that:

a) Some children, including those in crisis or who are particularly vulnerable, are more likely to be encouraged or instructed by some recovery content regardless of the intention behind the post. This could include detailed documentations of recovery journeys which could provide instructions for self-harm, or encouraging users to engage in ‘less extreme’ forms of self-harm. b) Harmful comments may be posted on genuine recovery content, for example, when users ‘take over’ another user’s post and flood it with harmful comments about self-harm. c) Harmful content about self-harm may be posted in genuine recovery forums, groups or communities.

4.9 In addition, providers should be aware that self-harm content develops quickly and evolves rapidly, which can enable such content to evade content moderation. This can include the use of code words and hashtags employed by users to signal and search for self-harm content. Often, these terms can be specific to a particular service. Some content may also adapt commonly used terms to obscure the content (e.g., intentional misspelling/leet,45 algospeak,46 use of acronyms, coded adoption of words originating in popular culture). Service providers should therefore keep in mind that while a post might not explicitly suggest self-harm, it could be accompanied by more suggestive comments or hashtags which indicate that the content does in fact meet the definition of self-harm content.

4.10 Service providers should take care when assessing artistic representations of suicide and self-harm, for example, music, poetry, artwork or fictional works, drawings, stories, paintings and other art. These may romanticise or glamourise self-harm in a way that is harmful to children. We are mindful of the need to consider freedom of expression. However, where content meets the definition of self-harm content under the Act, then it should be treated as self-harm content harmful to children, regardless of the format. Service providers should therefore focus on whether the content meets the definition under the Act.

Examples¶

Examples or kinds of content that Ofcom considers to be self-injury content that is harmful to children¶

4.11 Table 4.2 describes the kinds of self-harm content that Ofcom considers to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

4.12 We have not addressed the level or impact of harm within Table 4.2. For further information on the risk and impact of self-harm content, see Section 3 of the Children’s Register.

4.13 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

45 Leet speak or ‘l337 speak’ refers to an informal online language where numbers or special characters are used to replace vowels or consonants. 46 Algorithm speak or ‘algospeak’ refers to coded language used online to circumvent content moderation methods. This could also include the use of emojis.

4.14 Service providers should keep in mind that, in accordance with the Act, the kinds or types of content listed must encourage, promote or provide instructions for a deliberate act of self-injury, but do not need to do so intentionally.

Table 4.2: Descriptions and examples of content which encourages, promotes or provides instructions for an act of deliberate self-injury

Content description

Content examples

Content describing methods of self-harm

• A post describing a novel or uncommon method of

self-harm.

• A post describing a particular method of self-harm as

easy to do, easy to conceal, quick to heal or unlikely to

scar.

Content containing detailed instructions for methods of self-harm

• A post describing or depicting detailed steps to carry out a particular method of self-harm, whether novel or well known. • A post providing details about specific tools, substances and accessories for self-harm and where or how to obtain them. • A post describing an individual’s personal account of how they have self-harmed, could self-harm or plan to self-harm.

Content, such as internet challenges or dares, instructing self-harm

• Online challenges that dare or otherwise encourage

others to engage in behaviour which may amount to

self-harm methods. For example, challenges

encouraging self-asphyxiation or encouraging the

application of harmful substances to burn the skin.

Content featuring real-life self-harm

• Images of self-harm wounds, presented without appropriate context. Such images may be accompanied by promotional text. • Content that shows any individual actively preparing to self-harm, e.g., livestreaming or pre-recorded. This could include discussion of detailed plans or preparing accessories/tools for self-harm. • Content of a person self-harming, e.g., livestreaming or pre-recorded.

Content which encourages competition in relation to self-harm, or more extreme behaviour related to self-harm

• Posts or comments which promote or encourage self-

harm by suggesting that self-harm is competitive or

that certain wounds or methods of self-harm are

insufficiently severe. Such posts often use coded

language.

Content description

Content examples

Content which represents self-harm in a way which romanticises, glamourises, glorifies or normalises it

• Content depicting common accessories for self-harm with sounds or hashtags known to indicate or be associated with self-harm content (e.g., a photo or drawing of scissors with a sound known to indicate self-harm content). • Content describing self-harm as a healthy way of coping with problems (e.g., poems or songs). • Content describing self-harm as a sophisticated or artistic response to problems (e.g., fanfiction posts).

Content description

Content examples

Brief descriptions or mentions of

• A post which discusses someone’s experience of self-harm without detail or commentary on the methods.

self-harming method where the

description is very limited and well

known

Content depicting self-harm healed scars incidentally

• Content depicting healed self-harm scars where the overall focus of the content itself is not about self-harm. • Images of an individual which shows some healed scarring on their body but is not the focus of the content.

Content about concealing healed

• A post which advises how to cover up healed self-harm scars, e.g., with make-up or tattoos.

scars in a recovery context (not

concealing active self-harm

wounds, accessories or behaviours)

Content description

Content examples

Content which promotes recovery and awareness of self-harm in a healthy, non-harmful way and/or which reduces stigma

• Discussions of healthy coping mechanisms. • Descriptions of signs that may indicate someone may be self-harming and how to assist them. • Explanations of how to talk about self-harm in a non-stigmatising and safe way. • Information about how and where to seek further help regarding self-harm. • Facts and statistics about self-harm (but not methods). • Individuals discussing their own recovery or seeking support. • A post which discusses an individual’s ‘lived experience’ of self-harm, which does not provide details, commentary or specific methods of self-harm.

Safety-promoting resources which provide healthy coping mechanisms or signpost to services for recovery

• Charity or recognised self-harm prevention bodies’

web pages or resources.

• NHS information pages, National Institute for Health

and Care Excellence (NICE) clinical guidelines and

other recognised medical resources.

• Resources which encourage individuals to seek help or

speak to someone, or refer individuals to places where

they can seek help.

Content which describes or discusses safe ways to deal with negative emotions, feelings of wanting to self-harm, safe alternatives and coping strategies

• Descriptions or discussions of ways to recognise triggers. • Descriptions or discussions of how to distract from the urge to self-harm, such as exercise, tidying up or holding ice cubes. • Descriptions or discussions of ways to delay acting on feelings of self-harm, such as waiting at least five minutes to see if the urge disappears (increasing the delay over time).47

Discussions or content focused on

• Charity and educational sites describing the risks and signs of self-harm.

research, advocacy and support

related to self-harm awareness and

prevention

Academic or educational articles

• Facts and statistics about self-harm which raise awareness and reduce stigma. • Articles which do not discuss methods or include instructions, but otherwise research prevalence, incidence and other statistics related to self-harm, or self-harm methods.

Examples or kinds of content Ofcom considers not to be self-injury content that is harmful to children¶

4.15 The following table describes the kinds of self-harm content that Ofcom does not consider to meet the definition of PPC in the Act, along with a non-exhaustive list of examples.

4.16 Although Ofcom does not consider the examples in the following table to meet the definition of PPC, the content may still be inappropriate for children in certain age groups.

4.17 Ofcom does not expect services to unduly restrict children’s access to genuinely safety-promoting, non-harmful information and resources relating to self-harm. Where content promotes recovery and awareness of self-harm and does not contain any of the content listed in Table 4.2, this may not be considered PPC.

Table 4.3: Descriptions and examples of content which is not content that encourages, promotes or provides instructions for an act of deliberate self-injury

47 Mind, 2024. Helping yourself if you self-harm. [accessed 7 February 2025].

Section 5 Guidance on eating disorder content¶

Warning: this section contains references to content that may be upsetting or distressing, including detailed discussion of eating disorders and eating disorder content. Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be content which encourages, promotes or provides instructions for an eating disorder, or behaviours associated with an eating disorder.48 This is referred to as eating disorder content.

Considerations for service providers¶

5.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).49 In this section, we outline examples of content that Ofcom considers to be, or not be, eating disorder content, a form of primary priority content that is harmful to children (PPC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

5.2 People who are posting or interacting with eating disorder content may be in vulnerable or difficult circumstances. We are conscious of the stigma which may be associated with eating disorders and that this may present a barrier to individuals seeking help. It is important for children to have the opportunity to access and share content about their mental health and we have been mindful of this when developing this guidance. We also expect service providers to review this guidance in conjunction with the eating disorder content in Section 4 of the Children’s Register of Risks (Children’s Register) to better understand the sensitivities around this content.

5.3 Ofcom does not expect service providers to unduly restrict children’s access to genuine safety-promoting, non-harmful information and resources relating to eating disorders.

Relevant illegal content¶

5.4 Content may be considered illegal where it intentionally encourages or assists someone in an act of serious self-harm, including psychiatric injury. This could include instances where individuals encourage or assist others in carrying out behaviours linked to eating disorders that amount to serious acts of self-harm.

5.5 We have provided comprehensive guidance on how service providers should make illegal content judgements in the Illegal Content Judgements Guidance (ICJG) of our December 2024 Statement on Protecting People from Illegal Harms Online, and service providers should refer to this to understand the criteria for illegality in full.

48 Section 61(5) of the Online Safety Act 2023 (the Act). 49 Section 53 of the Act.

Key terms¶

Table 5.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Eating disorder

Eating disorders are serious and complex mental illnesses which are

characterised by disordered eating behaviour and a preoccupation

with food, often accompanied by concerns with body weight or body

shape.50

Behaviours associated with eating disorders

A non-exhaustive, general list of behaviours commonly associated with eating disorders, such as: calorie restriction or fasting, excessive exercise, binge eating, compensatory or purging behaviours, avoidance or restriction of food, and concealing behaviours from friends and family.

Encourage

Content which recommends, advises, incites or induces eating

disorders, or behaviours associated with an eating disorder.

Encouragement could be unintentional or inadvertent. This can include

content which glamourises, glorifies, romanticises or normalises eating

disorders.

Promote

Content which publicises, endorses or portrays eating disorders or behaviours associated with an eating disorder as positive, e.g., as a healthy coping mechanism or ‘lifestyle choice’. Promotion could be unintentional or inadvertent. This can include content which glamourises, glorifies, romanticises or normalises eating disorders.

Provide instructions for

Content which is instructive in nature. This could detail ‘tips’,

‘methods’, ‘coaching’ or ‘guidance’ specifically related to eating

disorders or behaviours associated with eating disorders, regardless of

whether the content was intended to encourage such instructions to

be followed or not.

This can include content which describes eating disorders or

behaviours associated with eating disorders in sufficient detail that can

be emulated.

This can also include content which provides instructions on

concealing eating disorder symptoms or behaviours.

Additional context for service providers¶

5.6 In some instances, there may be overlaps between eating disorder content and other categories of content such as suicide and self-harm content (see Sections 3 and 4), bullying

50 Treasure, J., Duarte, T. A. and Schmidt, U., 2020. Eating disorders, Lancet, 395 (10227), pp.899-911. [accessed 25 March 2024]; World Health Organization, 2019. International Classification of Diseases, Eleventh Revision. [accessed 25 March 2024].

content (Section 7), harmful substances content (Section 9) and dangerous stunts and challenges content (Section 10). For example, this could include online challenges which dare children to engage in behaviours associated with eating disorders.

5.7 Service providers should be aware of the different types of behaviours associated with eating disorders, as this can then inform their judgements. For example, eating disorders can present differently across different genders, with between a quarter and a third of those affected by eating disorders being men and boys. Service providers should also be aware of eating disorder presentations which are lesser known or under-researched, such as binge-eating disorder. For more information see Section 4 of the Children’s Register.

5.8 Service providers should also be mindful where content relating to eating disorders provides safety-promoting resources, and should ensure that children are able to access resources from genuine support services or qualified experts. Children should be able to access responsible, health-promoting information about, and support for, eating disorders.

5.9 Relatedly, service providers should carefully consider recovery content when making judgements. Recovery content can provide sources of hope, support, connection, encouragement and motivation for challenging eating disorder behaviours.51 However, there is a broad range of content related to recovery, and service providers should be aware that even some content which is framed as ‘recovery focused’ can nevertheless be associated with risk of harm to children.52 For example: a) Some representations of eating disorder recovery may include characteristics which may be harmful for children, such as images of emaciation presented as a ‘before’ and ‘after’ in a recovery context.53 This may include content which lists the number of calories consumed, documents specific weights or which provides accounts of eating disorder behaviours which are detailed enough as to provide instructions.54 b) Harmful comments may be posted on positive recovery content. Harmful content may also be disguised as ‘recovery focused’ through the use of recovery hashtags, for example.

5.10 Service providers should also be aware that pro-eating disorder content can develop quickly and evolve rapidly, which can enable such content to evade content moderation. Content tagging and code words can be used by pro-eating disorder communities, often in the form of hashtags or captions. For example, some content may be disguised as focused on ‘healthy’ lifestyles or ‘pro-fitness’.55 Some content may use terms intended to obscure the

51 Au, E. S. and Cosh, S. M., 2022. Social media and eating disorder recovery: An exploration of Instagram recovery community users and their reasons for engagement, Eating Behaviors, 46. [accessed 25 March 2024]. 52 Ofcom, 2024. Experiences of children encountering online content relating to eating disorders, self-harm and suicide. [accessed 25 March 2024]. Subsequent references to this source throughout.

53 Greene, A. K., Norling, H. N., Brownstone, L. M., Roe, C., and Moody, S., 2023. Visions of recovery: a cross- diagnostic examination of eating disorder pro-recovery communities on TikTok, Journal of Eating Disorders, 11. [accessed 25 March 2024]. 54 Beat, n.d. Eating Disorder Media Guidelines. [accessed 25 March 2024]. 55 Ofcom, 2024. Experiences of children encountering online content relating to eating disorders, suicide and self harm.

content (e.g., intentional misspelling/leet56, algospeak57, use of acronyms, use of coded adoption of words originating in popular culture or dieting communities).58

5.11 Within the context of pro-eating disorder forums, groups and communities, images can be shared to incentivise eating disorders or behaviours associated with eating disorders. ‘Thinspiration’, or variations on this term (often using the suffix -spo), are often also used to encourage eating disorder behaviours, or motivate users towards extreme thinness.

5.12 Service providers should take care when assessing artistic representations of eating disorders or behaviours, for example, music, poetry, artwork or fictional works, drawings, stories, paintings and other art. These may romanticise or glamourise eating disorders in a way that is harmful to children. We are mindful of the need to consider freedom of expression. However, where content meets the definition of eating disorder content under the Act, then it should be treated as eating disorder content, regardless of the format. Services should therefore focus on whether the content meets the definition under the Act.

Examples¶

Examples or kinds of content that Ofcom considers to be eating disorder content that is harmful to children¶

5.13 Table 5.2 describes the kinds of eating disorder content that Ofcom considers to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

5.14 We have not addressed the level or impact of harm within Table 5.2. For further information on the risks and impact of harm, see Section 4 of the Children’s Register.

5.15 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

5.16 Service providers should keep in mind that, in accordance with the Act, the kinds or types of content listed must encourage, promote or provide instructions for eating disorders, but do not need to do so intentionally.

56 Leet speak or ‘l337 speak’ refers to an informal online language where numbers or special characters are used to replace vowels or consonants. 57 Algorithm speak or ‘algospeak’ refers to coded language used online to circumvent content moderation methods. This could also include the use of emojis. 58 Center for Countering Digital Hate, 2022. Deadly by Design. [accessed 17 February 2025].

Table 5.2: Descriptions and examples of content which encourages, promotes or provides instructions for eating disorders, or behaviours associated with eating disorders

Content description

Content examples

Content which glamourises, glorifies, romanticises or normalises eating disorders, including images or videos of emaciation with aspirational commentary, or presented in a way which presents emaciation as aspirational

• Images shared as motivation to encourage,

glamourise, romanticise or glorify extreme thinness

(e.g., ‘thinspiration’), such as through using aesthetic

filters or commentary presenting emaciation as

aspirational.

• ‘Body checks’ and/or images of extreme thinness

romanticising and showing protruding collarbones or

hipbones; protruding rib bones or flat or concave

stomachs; ‘thigh gaps’; or ‘before and after weight-

loss transformations’.

• Images or videos presenting extreme weight loss or

poses which emphasise dangerously low body weight.

These may be accompanied by specific weights, body

mass index (BMI) or measurements.

• Content which glamourises, glorifies, romanticises or

normalises dangerously low body weight, particularly

when accompanied by promotional text, sounds or

hashtags.

• Images accompanied by text-based ‘motivational’

quotes or poetry encouraging eating disorders, or tags

or references to known pro-eating disorder hashtags.

Content which includes instructions for carrying out methods or accessories for eating disorders or behaviours associated with eating disorders

• Content which provides instructions for any methods of extreme calorie restriction, purging, bingeing or increasing energy expenditure. • Instructions for concealing eating disorder behaviours from others (e.g., how to hide purging or bingeing to continue to engage in such behaviours, how to conceal food restriction or fasting, how to conceal weight loss or lack of weight gain when being weighed). • Instructions for purging methods, or compensatory behaviours (e.g., information on types of laxatives, where to buy them, dosage, when to take them and alternatives that may have the same effect).

Content description

Content examples

Content which promotes or encourages competition or challenges in relation to eating disorders or behaviours associated with eating disorders

• Content which glamourises, glorifies or frames

symptoms of eating disorders as measures of success

or encourages competition.

• Content which lists specific calorie counts, weights,

BMI or other measurements within the context of an

eating disorder.

• Content which uses criticism as an incentive to engage

in eating disorder behaviours. For example, users

requesting or giving ‘meanspo’ (i.e., ‘mean’ or critical

comments to motivate eating disorder behaviours).

• Content which encourages disgust or ridicule towards

bodies which do not fit a thin or muscular ‘ideal’ in

order to incentivise eating disorder behaviours.

• Content which shames, ridicules or frames some

eating disorder behaviours as not ‘extreme enough’.

• Challenges related to eating disorders, including

challenges which encourage others to engage in, e.g.,

extreme weight loss or gaining muscle through

behaviours which are disordered.

Content which encourages interactive behaviours associated with eating disorders, including in pro-eating disorder or recovery spaces

• Content involving groups, forums or group messages which promote eating disorder behaviours, e.g., requesting or offering to act as an ‘ana-buddy’, ‘pro-ana coach’59 or engage in group fasting. • Content encouraging others to share images of emaciation, and/or images of purging. • Sharing or requesting ‘tips’, ‘hacks’ or ‘motivation’ for behaviours associated with eating disorders, with users listing goal weights, current weights and/or lowest weights (e.g., in forum signatures or bio information). • Lists of ‘commandments’ for eating disorder lifestyles.

Content description

Content examples

Content which describes behaviours associated with eating disorders in sufficient detail as to provide instructions for eating disorders, including when inadvertent or presented in a recovery context

• Content including detail about specific amounts of

calories or food consumed per day within the context

of an eating disorder.

• Blogs including detailed accounts of extreme or

persistent calorie restriction without countervailing

context.

• Stories which include details about eating disorder

behaviours or instructions which can be emulated,

with the lack of countervailing context. This could

include detailed description of extreme and/or

persistent calorie restriction or extreme fasting (e.g.,

including specific weights).

Extreme or excessive dieting, weight loss, nutritional or fitness content

• Content which provides instructions for weight loss or fitness which is extreme and/or excessive and/or dangerous, e.g., posts promoting excessive exercise despite injury or fatigue. • Content which provides instructions for excessive dieting or nutrition, such as posts documenting extreme and or persistent calorie restriction, extreme fasting, group fasting or drinking water to curb hunger. • Content which promotes laxative misuse or manipulation of insulin dosage.

Content description

Content examples

Safety-promoting resources which outline healthy coping mechanisms or signpost to services for recovery

• Content which provides support resources and

encourages help-seeking (e.g., charity helplines).

• Content which promotes the message that recovery

from an eating disorder is possible with the right

treatment and support.

• NHS information pages, National Institute for Health

and Care Excellence (NICE) Guidelines and other

recognised medical resources.

Content focused on research, awareness, advocacy and support related to eating disorder prevention

• Content from charity or educational sites posted on user-to-user services which could, e.g., give information about spotting the signs of eating disorders, or signpost where/how to seek help for an eating disorder.

Content which includes descriptions of eating disorders and/or associated behaviours, including recovery content, without sufficient detail to provide instructions, and which do not otherwise encourage or promote

• A post which discusses an individual’s lived experience

of an eating disorder which is not sufficiently detailed

as to provide instructions for behaviours associated

with eating disorders (e.g., which do not list specific

lowest weights or calories consumed per day).

• Text-based motivational quotes encouraging recovery

from an eating disorder.

• Lists of reasons to recover from an eating disorder,

recovery goals or distractions.

• Artistic works which motivate or describe eating

disorder recovery, where they do not promote or

encourage eating disorders (e.g., they do not include

images of emaciation or content which glamourises

eating disorders).

Academic articles or studies that examine eating disorders

• Eating disorder research which does not provide instructions for behaviours associated with an eating disorder. • Research that does not provide information about where individuals may find instructions for behaviours associated with eating disorders (e.g., which do not mention forums or chatrooms that contain content which encourages, promotes or provides instructions for eating disorder behaviours).

Journalistic content that does not

• An article focused on the prevalence of eating disorders.

promote, encourage or provide

instructions for eating disorders

Content description

Content examples

Content involving eating behaviours

• Content related to religious fasting. • Content related to infrequent overeating or feasting during holidays or occasional celebrations.

which are due to religious or other

culturally sanctioned practices

Content involving images of people with low body weight which do not glamourise or otherwise promote eating disorders

• Images of people with low body weight which is

unrelated to an eating disorder, e.g., due to another

medical condition.

• Content depicting people with low body weight or low

BMIs which is not presented in a manner or context

which glamourises or promotes eating disorders.

59 The terms ‘ana-buddy’ and ‘pro-ana coach’ are used to describe an individual who encourages or assists a person to maintain behaviours associated with an eating disorder and promotes an anti-recovery mindset.

Examples or kinds of content that Ofcom considers not to be eating disorder content that is harmful to children¶

5.17 Table 5.3 describes the kinds of eating disorder content that Ofcom does not consider to meet the definition of PPC under the Act, along with a non-exhaustive list of examples.

5.18 Although Ofcom does not consider the examples in the following table to meet the definition of PPC, the content may still be inappropriate for children of certain age groups.

5.19 Ofcom does not expect services to unduly restrict children’s access to genuinely safety-promoting, non-harmful information and support resources relating to eating disorders. Where content that promotes recovery and awareness of eating disorders, and does not contain any of the content listed in Table 5.2, it may not be considered PPC.

Table 5.3: Descriptions and examples of content which is not content that encourages, promotes or provides instructions for eating disorders, or behaviours associated with eating disorders

Section 6 Guidance on abuse and hate content¶

Warning: this section contains references to content that may be upsetting or distressing, including reference to sexual violence Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be content which is abusive and content which incites hatred. Content which is abusive60 means that which targets the characteristics of race,61 religion,62 sex, sexual orientation, disability63 or gender reassignment.64 65 Content which incites hatred66 means that which incites hatred against people of a particular race, religion, sex or sexual orientation; who have a disability; or who have the characteristic of gender reassignment.

Considerations for service providers¶

6.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).67 In this section, we outline examples of content that Ofcom considers to be, or not be, abuse and hate content, a form of priority content that is harmful to children (PC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

6.2 This guidance deals with abusive content and content that incites hatred jointly, as both types of content target listed characteristics and have similarities due to how these harms manifest. Throughout this guidance we refer to the two categories of content collectively as ‘abuse and hate content’. We refer to the characteristics outlined in the Act as ‘listed characteristics’ for the purposes of this guidance.

60 Section 62(2) of the Online Safety Act 2023 (the Act). 61 Section 62(10)(b) of the Act explains ‘race’ to include colour, nationality, and ethnic or national origin. 62 Section 62(10)(c) of the Act explains that references to religion include references to a lack of religion. 63 Section 62(10)(a) of the Act explains ‘disability’ to mean any physical or mental impairment. 64 Section 62(11) of the Act explains that a person has the characteristic of gender reassignment if the person is proposing to undergo, is undergoing or has undergone a process (or part of a process) for the purpose of reassigning the person’s sex by changing physiological or other attributes of sex.

65 Section 62(2) of the Act. We note that these characteristics are similar to certain ‘protected characteristics’ under the Equality Act 2010, which also lists disability, gender reassignment, religion (in that case, including belief), sex and sexual orientation among the protected characteristics. However, the way these terms are defined in the Act is not identical to the definitions used in the Equality Act 2010, so we do not presume they should have precisely the same meaning for the purposes of this guidance. 66 Section 62(3) of the Act. 67 Section 53 of the Act.

6.3 Although treated jointly, there are distinguishing factors between the two types of content. For the purposes of this guidance, we define abusive content as being targeted at the listed characteristics of an individual, whereas content that incites hatred is against a group, or a group of persons holding the same listed characteristic(s). This could be in the form of direct calls to hate a particular group, or where hateful expressions in themselves could have the effect of inciting others.

Relevant illegal content¶

6.4 Abuse and hate content that is targeted against some of the listed characteristics may be illegal in some circumstances. We have provided comprehensive guidance on how service providers should make illegal content judgements in relation to abuse, harassment and hate content in the ‘Threats, abuse and harassment (including hate)’ section of the Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online. Service providers should refer to this to understand the criteria for illegality in full. Broadly, content may be illegal where there are reasonable grounds to infer that it includes threats, abuse and insults (including hate) encompassing: a) threatening or abusive behaviour which is likely to cause a ‘reasonable person’ to suffer 'fear or alarm’;68 b) threatening or abusive behaviour which is likely to cause harassment or distress;69 c) threats or abuse which are likely to ‘stir up’ racial hatred;70 d) threats which are likely to ‘stir up’ hatred on the basis of religion or sexual orientation;71 e) threats or abuse which may provoke violence; f) other content likely to amount to harassment (including stalking and controlling or coercive behaviour).

Key terms¶

Table 6.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Abusive

Content that insults, derogates, dehumanises or threatens a person on

the basis of a listed characteristic.

Incites hatred

Content that encourages, advocates or justifies animosity, hostility, rejection or violence against a group of persons on the basis of a listed characteristic.

68 This is likely to amount to an offence under section 38 of the Criminal Justice and Licensing (Scotland) Act 2010. 69 This is likely to amount to an offence under section 5 of the Public Order Act 1986. 70 Racial hatred is hatred against a group of persons defined by reference to colour, race, nationality (including citizenship), or ethnic or national origins. This is likely to amount to an offence under section 21 of the Public Order Act 1986. 71 This is likely to amount to an offence under section 29B of the Public Order Act 1986.

Additional context for service providers¶

6.5 Abuse and hate content is a broad category and may overlap with other categories of content. For example, content which incites hatred against women can draw on themes promoting or encouraging gendered violence (see Section 8). It can also overlap with bullying content, for example, where offensive content is persistently or repeatedly targeted at people or groups based on certain characteristics (see Section 7).

6.6 In addition, service providers should be aware that content may be abusive or incite hatred on the basis of multiple listed characteristics at once. For example, abusive content may target a person on the basis of their race and sex, and this content may use different terms or be leveraged in different contexts than abusive content which targets either race or sex. Service providers should also note that abuse and hate content can be present regardless of whether the individual or group targeted holds the listed characteristic, or the individual or group is merely perceived to hold the listed characteristic.

6.7 Service providers should consider contextual factors when assessing potential abuse and hate content. Whether content is likely to be abusive or inciting hatred will depend on how the listed characteristic is presented, the negative stereotypes that exist and the social context. Service providers should consider the likelihood of abuse being experienced by an individual, or the likelihood of incitement of hate against a group, rather than the intent of the user posting the content. This means service providers should consider how and where the content is shared, who it is directed at and how it impacts the users exposed to it.

6.8 A wide range of actors engage in abusive and hateful behaviour, including both lone individuals and individuals affiliated with hate groups or groups advocating violence and hatred against protected groups with listed characteristics. Some actors may deliberately leverage online spaces in order to humiliate or degrade a target, and this may form part of a campaign of mistreatment. Such behaviour may also take the form of pile-ons, where groups of coordinated perpetrators target a specific individual or groups, often with abuse. While pile-ons can happen to any user, they often target individuals in public life, for example, female journalists or politicians.

6.9 In practice, substitute terms or symbols for listed characteristics might be used to evade content moderation. Such proxies can be neutral or have derogatory intentions. For instance, these might appear as hashtags as part of the content. Content might also be posted as a ‘one-off’ or be repetitive in nature. This could also indicate that the abuse and hate content overlaps with Section 7 (Guidance on bullying content), specifically where content is targeted at a listed characteristic.

Examples¶

Examples or kinds of content that Ofcom considers to be abuse and hate content that is harmful to children¶

6.10 Tables 6.2 and 6.3 describe the kinds of content that Ofcom considers to meet the definition of PC under the Act, along with a non-exhaustive list of examples. Where listed characteristics are described, this is for illustrative purposes, and it is not intended to give undue weight to any one characteristic.

6.11 We have not addressed the level or impact of harm within the tables. For further information on the risks and impact of abuse and hate content, see Section 5 in the Children’s Register of Risks.

6.12 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the tables. We have separated the two types of abuse and hate content into separate tables under the relevant sub-headings of ‘content that is abusive and harmful to children’ and ‘content that incites hatred and is harmful to children’.

Table 6.2: Descriptions and examples of content that is abusive and harmful to children¶

Content description

Content examples

Insulting or intimidating remarks or harmful stereotypes targeted towards an individual

• A comment that leverages harmful tropes to derogate

a person on the basis of their religion and/or race.

• A post that uses racial, religious, ableist, transphobic,

gendered and/or homophobic slurs to degrade a

person.

• A comment that intentionally misgenders a person

with the intention to humiliate, insult, offend or ‘out’

someone.

• A hateful or aggressive misogynistic comment

targeting a woman or girl.

Content which objectifies and/or demeans a person on the basis of their listed characteristic

• A post that claims an individual is physically or

mentally inferior or deficient on the basis of one or

more listed characteristic(s).

• A derogatory meme or caricature of a person, with

threatening, abusive, hurtful or harmful commentary

added.

• A post or comment attacking someone based on their

gender72 using offensive, demeaning language to

describe them.

• A post or comment demeaning someone because of a

mental or physical disability.

• A ‘skin’ applied to a character or avatar in a video

game that is uploaded by a user and can be seen by

other users, which portrays a demeaning or harmful

stereotype of a listed characteristic.

• Abusive, degrading, or derogatory comments made by

a user against other user(s) in a video game, based on

one or more listed characteristics.

72 We use the term ‘gender’ to refer to a child’s sex and to gender reassignment. In Section 62(11) of the Act, the characteristic of gender reassignment is defined as follows: “if the person is proposing to undergo, is undergoing or has undergone a process (or part of a process) for the purpose of reassigning the person’s sex by changing physiological or other attributes of sex.”

Table 6.3: Descriptions and examples of content that incites hatred and is harmful to children¶

Content description

Content examples

Repeating harmful and discriminatory ideas about another group in order to encourage others to share such beliefs

• A comment which justifies or promotes the social

exclusion of a group that shares a listed characteristic.

• A post which encourages the dehumanisation of a

group sharing a listed characteristic by comparing

them to non-humans.

• A post which normalises, encourages or justifies

extreme misogynistic attitudes and behaviours, such

as expressing views that women and girls should be

subjugated by men.

• A post which promotes the idea that being

transgender is not a valid identity.

• A post or comment expressing attitudes that one race

or ethnicity is superior to another.

Defending or legitimising threatening action against a group of people

• A comment that justifies hate against groups with one or more listed characteristics. • A post that condones behaviours or attitudes that dehumanise a group belonging to a listed characteristic. • A video that explicitly suggests that hateful behaviours towards a group sharing a listed characteristic leads to social status, wealth and fame.

Content description

Content examples

Educational content and personal accounts of abuse or hatred

• An educational history clip where stereotypical tropes

are implicitly or explicitly challenged or criticised.

• A personal testimony of a victim or survivor of

targeted abuse explaining in detail what happened to

them.

Counter-speech

• A debate where harmful narratives receive credible and compelling alternative messages.

Artistic (including satirical) content

• A music video that uses reclaimed slurs or tropes. • A poem or image discussing oppression of a group in an attempt to raise awareness about injustice.

Journalistic content and democratic or political speech

• A journalistic account of an incident of hate and abuse (e.g., as recounted or filmed by a witness). • A comment on a post by a politician or political candidate who has a listed characteristic, where the comment expresses strong disagreement with a policy position, rather than abuse or hatred against the poster’s listed characteristic.

Examples or kinds of content that Ofcom considers not to be abuse and hate content that is harmful to children¶

6.13 Table 6.4 describes the kinds of abuse and hate content that Ofcom does not consider to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

6.14 The examples of content listed in the table do not meet the definition of PC, but they may still be inappropriate for children in certain age groups (e.g., younger children).

Table 6.4: Descriptions and examples of content that is not abuse or hate content that is harmful to children

Section 7 Guidance on bullying content¶

Warning: this section contains references to content that may be upsetting or distressing. Summary This section provides examples of the content or kinds of content that Ofcom considers to be, or considers not to be, bullying content. Bullying content includes content targeted against a person which conveys a serious threat, is humiliating or degrading, or forms part of a campaign of mistreatment.73

Considerations for service providers¶

7.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).74 In this section, we outline examples of content that Ofcom considers to be, or not be, bullying content, a form of priority content that is harmful to children (PC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

Relevant illegal content¶

7.2 Bullying content may be illegal in some circumstances. We have provided comprehensive guidance on how services should make illegal content judgements in relation to abuse and harassment content in our ‘Threats, abuse and harassment (including hate)’ section of the Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online. Service providers should refer to this to understand the criteria for illegality in full.

7.3 Broadly, content may be illegal where there are reasonable grounds to infer that a user is engaging in threatening or abusive behaviour, where that behaviour is within the hearing or sight of a person likely to be caused harassment, alarm or distress. For example, where individuals send threatening or abusive texts, voice messages, image or comments, or make threats to harm someone or their family (see ‘Threats, abuse and harassment (including hate)’ in the ICJG).

7.4 Bullying content can also be related to child sexual abuse and exploitation offences, for example, where child sexual abuse material is being created or shared as part of a campaign of mistreatment against an individual, or where someone threatens to disclose an intimate image of a child. This includes self-generated indecent images as well as images of the child created using generative artificial intelligence technology (see ‘Child sexual abuse material’ in the ICJG).75 Where sexual messages are sent to a child in the context of bullying this may also amount to grooming (see 'Grooming' in the ICJG).

73 Section 62(12) of the Online Safety Act 2023. 74 Section 53 of the Act. 75 Self-generated imagery may include content created using generative artificial intelligence (GenAI) technologies, for example. See the ICJG for more details.

Key terms¶

Table 7.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Serious threat

To express in earnest an intent to inflict serious harm (e.g., to inflict

physical harm, reputational damage or other hostile actions on

someone).

Humiliating

An act that lowers an individual’s dignity or self-respect, embarrasses or demeans them, or otherwise subjects them to humiliation.

Degrading

An act that lowers an individual’s estimation, debases them or brings

them into disrepute.

Within the context of bullying content, this could include sending,

sharing or posting untrue messages or rumours about a victim which

may make them feel humiliated or damage their reputation or

friendships.

Campaign of

Repeated, persistent and intentional ill-treatment of an individual or a group.

mistreatment

Bullying

Bullying encompasses behaviour that has most, or all, of the following

traits:

• Persistent or repetitive.

• Intended to hurt someone (physically, verbally and/or

psychologically).

• Involves a perceived power imbalance (including, e.g.,

imbalances in physicality and/or perceived social status).

Bullying can occur in person or online, or both. Where it occurs online,

it is often referred to as ‘cyberbullying’.

Additional context for service providers¶

7.5 Bullying content can often be aimed at individuals from certain groups, for example, because of race, religion, sex, gender reassignment or sexual orientation. Bullying content may therefore overlap with other categories, for example, abuse and hate content (see Section 6). Bullying content may also overlap with content which encourages suicide or self-harm (see Sections 3 and 4), eating disorder content (see Section 5), or violent content (see Section 8). We recommend that service providers take these sections into account when considering content which is initially indicated to be bullying content.

7.6 As a harm, bullying is varied, subjective and highly contextual. Bullying can involve repeated or persistent behaviour which is intended to hurt a single person or multiple people.76

76 We consider that the definition of bullying ‘content’ under the Act is broad and may include, for example, any communications between users. It may also, in limited circumstances, extend to bullying behaviours where this manifests through user communications (subject to it meeting the definition of PC). See also Section 6: Bullying content in our Children’s Register.

Impacts are generally cumulative. Refer to Section 6 of the Children’s Register of Risks (Children’s Register) for more detail. The current section however provides guidance on identifying bullying content. For this kind of harmful content, contextual factors are particularly important when making judgements about whether content should be considered bullying content. In seeking to assess potential bullying content, service providers should consider reasonably available information in determining whether content constitutes bullying content. This may include information provided by a user as part of a complaint or through a service’s reporting process (e.g., user complaints against an abusive, controlling or coercive user).

7.7 Service providers should consider whether individual pieces of content should be viewed in the context of a wider campaign of mistreatment. For example, if a child is being repeatedly targeted by bullying content or targeted by several other users, this may indicate that the hurtful content shared should be considered bullying content. Such behaviour may also take the form of pile-ons, where groups of coordinated perpetrators target a specific individual or groups, often with abuse.

7.8 Service providers should also consider how to distinguish between bullying and behaviour that some may describe as ‘jokes’ or ‘banter’ between friends. For example, if the child being targeted is engaging or responding in a light-hearted way, or has otherwise positive interactions with the user, this may indicate that the content should not be considered bullying content.

7.9 This guidance focuses on bullying content covered by the Act. For user-to-user services, this means regulated user-generated content. For search services, this means search content. However, we acknowledge that some bullying involves exclusionary behaviours. For example, children can be removed or excluded from group chats as part of a bullying campaign. This does not involve direct communication between users. It also does not involve uploading, generating or sharing content. Such actions do not show up in search results. We therefore do not provide examples of these exclusionary behaviours in this guidance. However, they are important to consider when understanding the risk of harm from online bullying. Evidence of exclusionary behaviours is included in the Children’s Register (Section 6) and should be considered by service providers when conducting their risk assessments.

7.10 Due to the nature of bullying outlined above, some of the examples in this section are less specific than other sections in this guidance. This reflects the fact that it may be particularly important for service providers to consider contextual factors in identifying bullying content.

Examples¶

Examples or kinds of content that Ofcom considers to be bullying content that is harmful to children¶

7.11 Table 7.2 describes the kinds of bullying content that Ofcom considers to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

7.12 We have not addressed the level or impact of harm within Table 7.2. For further information on the risks and impact of bullying content, please see Section 6 of the Children’s Register.

7.13 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

Table 7.2: Descriptions and examples of content that is bullying content that is harmful to children¶

Content description

Content examples

Content that persistently or repetitively targets individuals or groups, and which is offensive or otherwise harmful – may include targeting people or groups based on certain characteristics, see also Guidance on abuse and hate content (Section 6)

• Persistent or repetitive offensive name calling. Name

calling can include words which are overtly offensive.

Commonly used words can also take an offensive

meaning (e.g., if related to a nickname that has been

given to someone with the purpose of humiliating

them).

• Persistently or repetitively sending messages,

comments, emojis or posts to, or about, an individual

that are likely to cause harm, including content which

humiliates, insults, intimidates, teases, offends or

otherwise hurts someone (e.g., comments about

someone’s attributes or personal circumstances,

physical appearance, body type, personality, family,

behaviours or actions).

• Persistently or repetitively sending messages,

comments, emojis or posts which are intended to

humiliate, insult, intimidate, tease, offend or

otherwise hurt a group of people or an individual (e.g.,

targeting a group because of their interests or

hobbies, including likes or preferences which are seen

to be immature; targeting a group of children for being

studious at school; or targeting a group because they

are perceived not to belong to the group that the user

or users belong to).

Content description

Content examples

Content depicting or relating to a specific individual in an offensive or otherwise harmful way, shared without their consent in order to humiliate

• Circulating images or videos, such as unflattering photos or video screengrabs, without the consent of the person in the image or video and with the purpose of humiliating them (e.g., a video of someone falling over accompanied with humiliating commentary). • Manipulated images which are derogatory to an individual (e.g., superimposing an image of a child into a context intended to be offensive). • Circulating recordings of an individual without their consent as a means of humiliation (e.g., personal voice notes, or recordings of someone showing distress). Sharing personal details or private/sensitive • information without permission, to humiliate, or as part of a campaign of mistreatment (i.e., doxxing).77 • A video, image or audio of someone being bullied. For example, a voice note of someone being subjected to name calling, or physical bullying such as hitting, slapping or pushing, in order to humiliate them. (Note: some of this may also constitute violent content, see Section 8).

Content contributing to group humiliation or harassment in online public spaces

• Comments on content or posts which exacerbate or

intensify the humiliation or degradation of an

individual. This could also include text posts

maliciously sharing of personal details or

private/sensitive information about an individual (i.e.,

doxxing or a ‘pile-on’78).

Serious threats or aggressive behaviours

• One-off or knee-jerk comments telling someone to harm themselves that do not amount to encouraging self-harm (e.g. ‘go jump off a bridge’). • Serious threats to release private/intimate information (i.e., doxxing). Comments or posts that indicate that a child is being • bullied or abused as part of an intimate relationship.

Content description

Content examples

Content that is not malicious, which

• A clip from a stand-up comedy sketch which includes satirical or parodical jokes which are not malicious (i.e., not humiliating or degrading an individual or group, and not part of a wider campaign of mistreatment towards an individual or group).

does not humiliate or degrade and

does not form part of a campaign of

mistreatment, even if it may be

capable of causing some offence to

an individual

Content which raises awareness of bullying

• Educational or journalistic content, e.g., a video or documentary clip which features real or acted bullying for the purposes of raising awareness of bullying or criticising bullying behaviours. • A post from a charity signposting to support.

Depictions or descriptions of

• Clips from age-appropriate TV or film depicting or

bullying in dramatic contexts

describing bullying.

Online Safety Act 2023 (the Act).82 In this section we outline examples of content that

Ofcom considers to be, or not be, violent content, a form of priority content that is harmful

to children (PC) under the Act. For the purposes of this guidance, all mentions of image,

video or audio content included in the tables refer to both real and synthetic images, videos

or audio content.

Examples or kinds of content that Ofcom considers not to be bullying content that is harmful to children¶

7.14 Table 7.3 describes the kinds of bullying content that Ofcom does not consider to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

77 ‘Doxxing’ is referred to as ‘private/intimate information made public’. While this on its own may not appear as bullying, it can be classed as such when it is part of a campaign of mistreatment of a person. 78 ‘Pile-on’ refers to when a user is criticised or targeted by a large number of other users, often as part of bullying campaigns.

7.15 Although Ofcom does not consider the examples in the table to meet the definition of priority content, the content may still be inappropriate for children in certain age groups.

7.16 Service providers should keep in mind that, in accordance with the Act, the kinds or types of content listed must not constitute bullying content. Some examples would only apply in certain limited circumstances and would be context specific, so services should rely on reasonably available information, including user complaints, in making content judgements.

Table 7.3: Descriptions and examples of content that is not bullying content that is harmful to children

Section 8 Guidance on violent content¶

Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be violent content. Violent content means content which encourages, promotes or provides instructions for an act of serious violence against a person,79 or content which depicts real or realistic serious violence against, or detailed, graphic injury of, a person,80 animal or fictional creature.81

Considerations for service providers¶

8.1 This guidance relates to duties for user-to-user and search services under Part 3 of the

Relevant illegal content¶

8.2 Violent content is a broad category that encompasses many types of content, some of which may be illegal in some circumstances.

8.3 We have provided comprehensive guidance on how service providers should make illegal content judgements in relation to various kinds of illegal harm involving violence in the Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online (December 2024 Statement). Service providers should refer to this to understand the criteria for illegality in full.

8.4 Broadly, content may be illegal where it: a) Involves depictions of sexual violence involving an adult (see ‘Image-based adult sexual offences’ in the ICJG); b) Calls on others to commit violence on the basis of race, religion or sexual orientation (see sub-sections ‘Threats which are likely to stir up racial hatred’ and ‘Threats which stir up religious hatred or hatred on grounds of sexual orientation’ under ‘Threats, abuse and harassment (including hate)’ in the ICJG); c) Calls on others to commit violence for the purposes of what may be terrorism, or on behalf of a proscribed group (see ‘Terrorism’ in the ICJG); d) Involves insulting or abusive words or behaviour, or threatening words, signs or other visible representations and intends to provoke the immediate use of violence by any person, or to make someone think that such violence may be provoked (see sub-section

79 Section 62(4) of the Online Safety Act 2023 (the Act). 80 Section 62(6) of the Act. 81 Section 62(7) of the Act. 82 Section 53 of the Act.

‘Abuse and insults which may provoke violence’ under ‘Threats, abuse and harassment (including hate)’ in the ICJG);

e) Concerns the sale of a firearm, knife or weapon (see ‘Firearms and other weapons’ in the ICJG); f) Amounts to coercive or controlling behaviour (CCB). Note that CCB offences are perpetrated in complex and personal ways and can include threats of physical or sexual violence, but also non-physical CCB (refer to Section 5 of our Illegal Harms Register of Risks (Illegal Harms Register) and in the ICJG Section 3 on ‘Threats, abuse and harassment (including hate)’ and ‘Risk assessment and illegal content relating to controlling and coercive behaviour’); g) Encourages, assists or conspires to cause unnecessary suffering to an animal (see ‘Animal cruelty’ in the ICJG); h) Depicts what appears to be the real and deliberate killing or serious injury of humans or animals for the purposes of entertainment or amusement (see sub-section ‘Improper use of a public electronic communications network offence’ under ‘Non-priority offences and relevant non-priority offences (“other” offences)’ in the ICJG); and/or i) Depicts what appears to be the real torture of humans and/or animals (see sub-section ‘Improper use of a public electronic communications network offence’ under ‘Non-priority offences and relevant non-priority offences (“other” offences)’ in ICJG).

8.5 Any content which depicts children aged under 18 engaged in, or appeared to be engaged in, sexual activity is child sexual abuse material, which is illegal content (see ‘Child sexual exploitation and abuse (CSEA): Offences relating to child sexual abuse material (CSAM)’ in the ICJG).

8.6 Material connected to child criminal exploitation may also amount to illegal content. Service providers can refer to our December 2024 Statement, where child criminal exploitation is covered, for example, in section 9 of the Illegal Harms Register of Risks on human trafficking at paragraphs 9.33-9.41 under the sub-heading ‘Criminal exploitation including county lines’. Our ICJG provides further support to service providers in making judgements on this type of content on their service.

Key terms¶

Table 8.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Encourages

Content which may persuade others to engage in serious violence

and/or makes others more likely to engage in an act of serious

violence.

Promotes

Content which recommends, endorses or portrays violence as positive. This can include content which glamourises, glorifies, romanticises or normalises serious violence.

Provides instructions for

Content which details methods or guidance specifically related to an

act of serious violence, regardless of whether the content was

intended to encourage such instructions to be followed or not.

Term

Definition

Serious violence or

That for which the risk is not slight or negligible. It may be likely to cause significant, lasting or permanent physical and/or mental harm to others.

injury

Realistic violence or

A depiction of serious violence or injury that is accurate or true to life.

injury

This includes accurate or true to life computer-generated imagery or

content created by generative artificial intelligence (GenAI).

Content description

Content examples

Content that promotes or glamourises the carrying of weapons

• A realistic image or video of individuals or groups

posing with or brandishing weapons such as lock-

backs (flick blades), knives, axes, guns, clubs,

screwdrivers and machetes, and which depicts the

possession or use of such weapons in a positive light.

Content which promotes, glamourises, glorifies, romanticises or normalises violence

• A post which justifies or defends the use of sexual violence. • A video which commends domestic abuse as a means to ‘control’ women. • A post or video which idolises past instances of violence, e.g., a post demonstrating support or approval of a school shooting. • An image which idolises perpetrators of past instances of violence, e.g., a post depicting a murderer as saint-like. • A video that condones violence against particular groups or individuals, or otherwise explicitly suggests that such violent behaviours or attitudes are associated with an aspirational lifestyle, with fame and/or wealth.

Content which trivialises or misrepresents violent acts, where the purpose is to normalise or discount the impact of violent behaviour

• A post which argues that victims and survivors of

sexual assault must bear some responsibility.

• A video which mocks a public figure who has

experienced a serious act of violence to discredit or

degrade them.

Content description

Content examples

Content that encourages a serious act of violence

• A post or video that encourages or provides instruction on how to commit a serious act of violence, e.g., by providing information on how to purchase knives online for the purpose of serious violence. • A post or video that uses incendiary language and/or visuals about an incident, such as the injury or killing of another individual, including to encourage a serious act of violence in retaliation. • A post or video that encourages serious violence against an individual or group, or encourages others to seek to commit an act of serious violence towards another individual or group. This can include language communicated via song lyrics.

Content description

Content examples

Content depicting physical fights between groups or individuals, with or without the use of weapons

• An individual being stamped on or punched/kicked

repeatedly.

• An individual being forcefully hit in the head during a

fight.

• An individual being stabbed with a knife during a fight.

Content depicting multiple people beating up an individual

• Videos depicting a group of people beating up an individual in graphic detail. • Videos depicting an individual being beaten up by a group as a form of retaliation or punishment.

Content depicting individuals being subjected to threats, intimidation or degradation

• A video of a group breaking into a house and forcing

the inhabitants to perform humiliating acts under

duress.

• A video of an individual being threatened or physically

hurt while being made to beg and apologise.

Content depicting serious injury of a person in graphic detail, and often including blood and gore

• Images of a serious stab wound. • A video of the aftermath of a fight where an individual has been stabbed or otherwise seriously hurt, where cuts and damage to the skin and tissue are visible. • Videos or images of the aftermath of a car accident where an individual’s serious injuries are visible, such as a lot of visible blood, deep cuts or body mutilation.

Content description

Content examples

Content depicting serious violence and/or injury of a person in the context of war and other conflicts

• Depictions of serious violence and injury in graphic

detail without adequate contextualisation.

• Detailed images of the cracked skulls of individuals.

• Videos or images of the aftermath of a bomb

explosion or violent attack where the impact on

human bodies is visible in graphic detail.

Gaming content depicting serious violence or injury against a realistic human character

• A livestream or clip of a video game posted online

where a detailed, realistic CGI (computer-generated

imagery) human character is subjected to severe

physical and/or sexual violence in graphic detail.

• A clip of a video game posted online where a detailed,

realistic CGI human character is exploded.

Content description

Content examples

Content depicting an act of physical

• A video of an animal being beaten.

violence towards an animal

• A video depicting the torture of an animal.

Content depicting activities that

• A video of an organised hunt where animals are used to attack or kill other animals. • Videos or images showing organised animal fights.

involve the injury or death of

animals including animal hunts

organised by humans and staged

fights

Content depicting the serious injury

• Videos or images depicting the graphic or detailed aftermath of an animal hunt. • Videos or images depicting the graphic or detailed aftermath of an animal fight.

of an animal which is not

appropriately contextualised as a

natural occurrence. The injury is in

graphic detail, and/or includes

blood and gore

Content depicting accidental,

• Videos or images depicting an animal suffering from extreme starvation. • Videos or images depicting an animal suffering from living in extremely neglectful conditions.

negligent or abusive acts that have

resulted in the serious injury of an

animal without appropriate context

(e.g., which might ‘normalise’

suffering)

Gaming content depicting serious realistic violence or injury against an animal or fictional creature

• A clip of a video game posted online where a detailed,

realistic CGI animal is ripped apart.

• A livestream of a game where a detailed, realistic CGI

fictional creature is decapitated.

• A clip of a video game posted online where a detailed,

realistic CGI animal is exploded.

Content description

Content examples

Content that depicts or references violence but does not actively encourage it

• A video with language or visuals referencing knife

crime without glamourising it.

• A video depicting people with weapons without

encouragement or promotion of those weapons.

Journalistic, political or educational content

• A journalistic video discussing youth violence. • A post about an academic study on youth violence. • A post about the perceived failure of politicians or the police to tackle violence or intervene in a war, which gives (non-graphic) examples of real-life violence as context. • A testimony of a victim or survivor of violence explaining their experience.

Content description

Content examples

Content depicting violence that is clearly intended for humour or entertainment

• A video depicting slapstick violence where the tone is

exaggerated, cheerful or light-hearted.

• A video of an individual hitting another who falls

without sustaining serious injury.

• A video of a comedic knock to the head that lacks

force and does not cause serious injury.

• A video featuring playful physical contact, e.g.,

someone tackling another individual without force.

• An image depicting a fictional character carrying a

weapon, e.g., a cartoon, fantasy or fictional character

with a sword or axe.

Content depicting physical fights that would not likely meet the threshold for serious violence

• A video of a fight where individuals are pushing and shoving each other. • A clip from an age-appropriate film showing a fictional fight.

Content depicting superficial

• Images depicting scratches, bruises or cuts from the aftermath of a fight.

injuries where no lasting physical

harm is shown or implied

Journalistic content or content with an educational purpose

• Videos of individuals or groups in wartime conflict zones suffering from serious injuries as part of a journalistic article. • Videos or images of individuals or groups suffering from serious injuries due to a natural disaster or effect of the climate crisis as part of a journalistic article. • Videos or images from a warzone where violence is not shown in graphic detail, e.g., it is blurred or captured from a distance. • Historical depictions of violence or weaponry, where violence is not shown in graphic detail, e.g., battle re-enactments.

Content depicting professional or licensed sports

• A video of professional boxing, wrestling or martial

arts.

• A video of rugby tackling.

Gaming content depicting trivial

• A livestream of an age-appropriate video game where a detailed, realistic CGI human character is punched.

violence/injury against a realistic

human character

Content description

Content examples

Cartoon content depicting violence that is clearly intended for humour or entertainment

• A meme of a cartoon animal being eaten by another

animal.

• A video clip of a cartoon animal dying or being killed.

• Content that incorporates clips from age-appropriate

TV/film entertainment, e.g., popular cartoons with

animal characters.

Content that is journalistic or has

• A journalistic video discussing animal cruelty. • A video clip from a nature documentary showing an animal predator hunting prey.

educational purposes or

justification

Charitable or animal welfare content

• A video of an animal in poor living conditions,

appealing for further information (such as about how

the animal came to be living in such conditions).

• Images of animals in poor living conditions being used

to fundraise.

• A video of animals in poor living conditions as part of

an animal cruelty awareness raising campaign.

Gaming content depicting trivial

• A clip from an age-appropriate game posted online that depicts a fictional creature being playfully hit on the head.

violence/injury against a realistic

animal or fictional creature

Additional context for service providers¶

8.7 Violent content is a broad category and may overlap with other kinds of primary priority content that is harmful to children or PC. For example, content encouraging violence against women and girls may overlap with abuse and hate content (see Section 6) where content is abusive or incites hatred and is targeting characteristics of sex, sexual orientation, or gender reassignment.

8.8 Some forms of violent content, in particular content which encourages or promotes violence, may be more difficult for service providers to identify. For example, understanding the degree to which the content encourages or promotes violence can be reliant on an understanding of local events, social dynamics or identities. Service providers should be mindful of context when assessing whether content encourages or promotes violence.

8.9 Depictions of violence can be present in journalistic content83 or content of democratic importance, for example, in citizen journalism or advocacy during war. It is important that children are given the opportunity to access this content where it is appropriately contextualised, especially if violence is blurred or shot from a distance. However, where violence is graphic and depicts serious injuries without suitable context, it may be harmful to children, even if the content is journalistic or of democratic importance.

8.10 Many children play games online and can encounter violent content on gaming services, including via online gameplay and through messaging functionalities. Children can also encounter violent gaming content on other regulated services, for example, violent gameplay being livestreamed, or posted online. In assessing whether violent gameplay content poses a risk of harm to children, service providers should consider whether the violence depicted meets the definitions outlined in Table 8.1. Namely, whether the content depicts realistic serious violence or realistic injury in graphic detail. In making this assessment, it may be useful to refer to games industry age ratings.

Examples¶

Examples or kinds of content that Ofcom considers to be violent content that is harmful to children¶

8.11 Tables 8.2-8.4 describe the kinds of violent content that is harmful to children that Ofcom considers to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

83 See Section 1 for more information on journalistic content.

8.12 We have not addressed the level or impact of harm within the table. For further information on the impact of harm, please see the Children’s Register of Risks (Section 7).

8.13 Many kinds of content will only meet the definition under the Act in certain circumstances, which we illustrate with examples in the tables.

8.14 We have separated the three types of violent content into tables under the relevant subheadings as follows:

i Content which encourages, promotes or provides instructions for an act of serious violence against a person (Table 8.2);

ii Content which depicts real or realistic serious violence against a person, or the serious injury of a person in graphic detail (Table 8.3); and

iii Content depicting real or realistic serious violence against an animal or fictional creature, or the serious injury of an animal or fictional creature in graphic detail (Table 8.4).

Table 8.2: Description and examples of content that encourages, promotes or provides instructions for an act of serious violence against a person

Table 8.3: Description and examples of content that depicts real or realistic serious violence against a person, or the serious injury of a person in graphic detail

Table 8.4: Depicts real or realistic violence against an animal or fictional creature, or the serious injury of an animal or fictional creature in graphic detail

Examples or kinds of content that Ofcom considers not to be violent content that is harmful to children¶

8.15 Tables 8.5-8.7 describe the kind of content that Ofcom does not consider to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

8.16 The examples of content listed below do not meet the definition of PC, but they may still be inappropriate for children in certain age groups.

8.17 As above, we have separated out the three types of violent content into tables under the relevant subheadings as follows:

i Content which does not encourage, promote or provide instructions for an act of serious violence against a person (Table 8.5);

ii Content which does not depict real or realistic serious violence against a person, or the serious injury of a person in graphic detail (Table 8.6); and

iii Content which does not depict real or realistic serious violence against an animal or fictional creature, or the serious injury of an animal or fictional creature in graphic detail (Table 8.7).

Table 8.5: Content which does not encourage, promote or provide instructions for an act of serious violence against a person

Table 8.6: Description and examples of content which does not depict real or realistic serious violence against a person, or the serious injury of a person in graphic detail

Table 8.7: Descriptions and examples of content which does not depict real or realistic serious violence against an animal or fictional creature, or the serious injury of an animal or fictional creature in graphic detail

Section 9 Guidance on harmful substances content¶

Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be harmful substances content. ‘Harmful substances content’ means content which encourages a person to ingest, inject, inhale or in any other way self-administer (i) a physically harmful substance, or (ii) a substance in such a quantity as to be physically harmful.84

Considerations for services¶

9.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).85 In this section, we outline examples of content that Ofcom considers to be, or not be harmful substances content, a form of priority content that is harmful to children (PC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

Relevant illegal content¶

9.2 Encouraging or providing the means for a person to ingest, inject, inhale or self-administer a physically harmful substance may be illegal in some circumstances. We have provided comprehensive guidance on how services should make illegal content judgements on content related to illegal or harmful substances in our Illegal Content Judgements Guidance (ICJG), published as part of our December 2024 Statement on Protecting People from Illegal Harms Online (December 2024 Statement). Services should refer to this to understand the criteria for illegality in full. Broadly, content may be illegal where there are reasonable grounds to infer that it: a) Amounts to an offer to supply a ‘controlled drug’ or ‘psychoactive substance’, for example, because it offers either substance for sale (see Section 7 in the ICJG); or b) Intentionally assists or encourages an act of suicide through the administration or consumption of a physically harmful substance (see Section 13 in the ICJG); or c) Intentionally assists or encourages an act of serious self-harm (i.e., self-harm which would result in grievous bodily injury or severe injury, including psychiatric harm) through the administration or consumption of a physically harmful substance once or repeatedly (see the ‘Encouraging or assisting serious self-harm’ offence in Section 16 of the ICJG).

84 Section 62(9) of the Online Safety Act 2023. 85 Section 53 of the Act.

Key terms¶

Table 9.1: Terms frequently used in this section of the guidance or terms listed in Act¶

Term

Definition

Harmful substances

Substances that would cause physical harm to a person if ingested,

injected, inhaled or otherwise self-administered (regardless of

whether or not it is legal to sell or buy the substance in the

circumstances in question). This can include substances that may

cause physical harm to a person if ingested, injected, inhaled or self-

administered in any other way as a result of consuming it in excessive

quantities, or in ways which contravene legal or medical protocols for

its safe consumption.

We consider physically harmful substances, or substances that may be

harmful when used to excess, to include:

• Illegal drugs or psychoactive substances.

• Alcohol, e-cigarettes or tobacco products (which it is not legal to

sell to children in the UK due to the fact they could cause them

harm).

• Unregulated medicines, consuming prescription medicines without

a prescription, and non-prescription medicines in excess.

• Medical or cosmetic substances or treatments administered

contrary to regulation.

• Substances not intended for human consumption.

Encourages

Content which recommends, advises, incites, induces or instigates children to ingest, inject, inhale or in any way self-administer a physically harmful substance or a substance in such a quantity as to be harmful. For harmful substances specifically, we consider ‘encourage’ to mean active or express encouragement. Harmful substances content does not need to intentionally encourage acts highly likely to result in serious injury or death in order to be harmful to children. This is different to the illegal content described in paragraph 9.2 b) and c) where intent is needed to infer illegality.

Additional context for service providers¶

9.3 Service providers should be mindful that there may be overlap with other categories of harmful content, such as ingestion challenges that encourage individuals to ingest harmful substances (see Section 10). There may also be an overlap with suicide or self-harm content (see Sections 3 and 4) as well as eating disorder content (see Section 5).

9.4 Harmful substances content can be identified by specific features. We consider harmful substances content to actively or expressly encourage use of a particular substance. This is likely to include a recommendation or call to action. Content that depicts or describes the use of a harmful substance, without actively or expressly encouraging use, is unlikely to be in scope of the definition. For example, content may depict people using harmful substances in educational or dramatic contexts, in which there is no active or express

encouragement to use that substance. However, examples of content depicting substance use that do not meet the definition of harmful substances content set out in the Act may still be inappropriate for children in certain age groups. The British Board of Film Classification (BBFC) provides additional guidance for contexts in which depictions and descriptions of substance use in video works may be unsuitable for children in different age groups.86

9.5 Content which offers harmful substances such as alcohol and tobacco products for sale must meet the definition of user-generated content or search content under the Act in order to be in scope of the Act. For example, where a user is remunerated for promoting alcohol or tobacco, such as sponsored content shared by influencers, it would amount to user-generated content, and therefore may be considered harmful substances content.

9.6 For some harmful substances, there is limited evidence of examples that actively or expressly encourage use beyond paid-for advertising (which is subject to advertising regulation). Tobacco products are an example of this. We therefore have not included examples relating to tobacco products.

Examples¶

Examples or kinds of content that Ofcom considers to be harmful substance content that is harmful to children¶

9.7 Table 9.2 describes the kinds of harmful substances content that Ofcom considers to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

9.8 We have not addressed the level or impact of harm within Table 9.2. For further information on the risks and impact of harmful substances content, please see Section 8 of the Children’s Register of Risks.

9.9 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

Table 9.2: Descriptions and examples of harmful substances content that is harmful to children¶

Content description

Content examples

Content encouraging consumption of illegal drugs and psychoactive substances

• Depictions or personal accounts of users’ ingesting a

drug or psychoactive substance with a call to action

for others to try it.

• Livestreams of a person using a substance,

highlighting positive experiences such as the ‘high’

after consumption.

Content description

Content examples

Content encouraging consumption of unregulated medicines, prescription medicines without a prescription, and non-prescription medicines in excess. This often includes instructing the use of a harmful substance to achieve a certain outcome

• Content promoting the abuse of prescription medication, e.g., encouraging use of an anti-anxiety drug for recreational use, or a drug to treat attention deficit hyperactivity disorder (ADHD) as a ‘study drug’, or in quantities in excess of the recommended dose. • Content which encourages taking quantities of non-prescription medication (e.g., paracetamol, ibuprofen) in excess of the recommended dose. • Content encouraging the abuse of steroids or steroid-like drugs, including videos with first-hand accounts of the use of these substances which present this behaviour in a positive light or advocate others’ use as part of a fitness regime. • Content encouraging self-administered abortion methods or abortion ‘reversal’ remedies that involve ingesting, injecting, inhaling or in any other way self-administering a harmful substance.87

Content encouraging excessive consumption of alcohol, or consumption of other age-restricted substances

• Content encouraging activities that involve drinking

excessive amounts, often to the point of vomiting or

passing out, e.g., encouraging initiations or hazing

rituals or other challenges involving excessive

consumption of alcohol.

• Content encouraging use of e-cigarettes, e.g., a

sponsored post by an influencer promoting an e-

cigarette brand or product.

Content encouraging consumption

• Content encouraging the consumption of household cleaning materials as a medical cure, e.g. consuming a disinfectant to kill the Covid-19 virus.

of harmful substances not intended

for human consumption

Content encouraging use of

• Content encouraging or providing instructions for self-

cosmetic treatments which can

administration of cosmetic treatments, e.g., Botox

present risk of harm if not

injections or lip filler, that is contrary to regulation of

administered by qualified

those substances.

professionals

Content encouraging challenges or

• Ingestion challenges which might involve harmful substances, or substances in quantities that are physically harmful, e.g., cleaning products or non-prescription medication.

stunts that involve consumption of

harmful substances: see also

Guidance on dangerous stunts and

challenges content (Section 10)

Content description

Content examples

Journalistic content describing, depicting or referencing the use of harmful substances

• Journalistic content debunking or exposing false

claims around medical treatments involving harmful

substances (e.g., the efficacy of drinking disinfectant

as a cure for Covid-19).

• Journalistic content discussing ‘viral’ ingestion

challenges and the risks associated with them.

• Journalistic content on drug or alcohol misuse and its

relationship with social issues.

• Journalistic content on campaigns to legalise certain

substances (e.g., reporting on a campaign to legalise

the consumption of cannabis in the UK).

Content which depicts use of substances for dramatic purposes, taking context and age-appropriateness into consideration

• Clips from age-appropriate films or TV shows which depict use of harmful substances.88 • Content depicting or describing individuals drinking alcohol, smoking tobacco-based products or using e-cigarettes, without active or express encouragement of others to do the same.

Appropriate medical advice from reputable or accredited resources

• Appropriate medical advice from official sources (e.g.,

NHS social media campaigns on drugs, alcohol,

cigarettes or other harmful substances).

Content describing or depicting the use of harmful substances for purpose of education or information

• Content describing or explaining the mental or physical effects of a harmful substance (e.g., on websites supporting those or providing information to those struggling with substance use). • Content explaining how harmful substances can be administered – for the purpose of educating children on recognising harmful substances or understanding the risks.

86 ‘Drugs’ in BBFC, 2024. Classification Guidelines. [accessed 21 January 2025]. Subsequent references to this source throughout.

87 This example may be illegal, amounting to encouraging serious self-harm. For further information, refer to Sections 13 and 16 of the Illegal Content Judgements Guidance in our December 2024 Statement on Protecting People from Illegal Harms Online. [accessed 5 February 2025].

Examples or kinds of content that Ofcom does not consider to be harmful substances content that is harmful to children¶

9.10 The following table describe the kinds of harmful substances content that Ofcom does not consider to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

9.11 Although Ofcom does not consider the examples in the following table to meet the definition of PC, the content may still be inappropriate for children in certain age groups.

Table 9.3: Descriptions and examples of content that is not harmful substances content that is harmful to children

88 Source: ‘Drugs’ in BBFC, 2024. Classification Guidelines.

Section 10 Guidance on dangerous stunts and challenges content¶

Summary This section provides examples of the content or kinds of content that Ofcom considers to be or considers not to be dangerous stunts and challenges content. Dangerous stunts and challenges content means content which encourages, promotes or provides instructions for a challenge or stunt highly likely to result in serious injury to the person who does it or to someone else.89

Considerations for services¶

10.1 This guidance relates to duties for user-to-user and search services under Part 3 of the Online Safety Act 2023 (the Act).90 In this section, we outline examples of content that Ofcom considers to be, or not be, stunts and challenges content, a form of priority content that is harmful to children (PC) under the Act. For the purpose of this guidance, all mentions of image, video or audio content included in the tables refer to both real and synthetic images, videos or audio content.

Relevant illegal content¶

10.2 Dangerous stunts and challenges content could result in serious self-harm or death. Dangerous stunts and challenges content may therefore be illegal in some circumstances. Broadly, content should be considered illegal where there are reasonable grounds to infer that it amounts to an offence of intentionally encouraging or assisting suicide or an act of serious self-harm (i.e., self-harm which would result in grievous bodily harm or severe injury, including psychiatric harm).

10.3 We have provided comprehensive guidance on how services should make illegal content judgements in connection with suicide and self-harm offences in Section 13 (Encouraging or assisting suicide) and Section 16 (Non-priority offences and relevant non-priority offences: sub-section: Encouraging or assisting serious self-harm) of our Illegal Content Judgements Guidance (ICJG) published as part of our December 2024 Statement on Protecting People from Illegal Harms Online (December 2024 Statement). Service providers should refer to this to understand the criteria for illegality in full.

89 Section 62(8) of the Online Safety Act 2023. 90 Section 53 of the Act.

Key terms¶

Table 10.1: Terms frequently used in this section of the guidance or terms listed in the Act¶

Term

Definition

Encourages

Content which recommends, advises, incites, induces or instigates

children to engage in challenges or stunts highly likely to result in

serious injury.

Dangerous stunts and challenges content does not need to

intentionally encourage acts highly likely to result in serious injury or

death in order to be harmful to children. This is different to the illegal

content described in paragraph 10.2-10.3 where intent is needed to

infer illegality.

Promotes

Content which actively publicises, endorses or portrays replicating/emulating stunts or challenges, that are highly likely to result in serious injury, as positive.

Provides instructions for

Content which is instructive in nature, including describing steps, tips

or methods to engage in challenges or stunts which are highly likely to

result in serious injury, in sufficient detail or clarity so that they may be

emulated.

Serious injury

Likely to cause significant, lasting or permanent physical harm or psychiatric harm.

Stunt

An unusual or difficult action that is done to attract attention, for

spectacle or entertainment purposes, to get publicity or popularity for

the individual, group or organisation responsible for it.91

Challenge

A difficult or demanding task, especially one seen as a test of one’s abilities, strength or character.92

Additional context for service providers¶

10.4 Service providers should also be mindful that there may be overlap with other categories of harmful content. Such content may include ingestion challenges that involve consumption of harmful substances (e.g., cleaning products), or of substances in such a quantity as to be physically harmful (e.g., challenges encouraging excessive consumption of medication or food products) (see Section 9). Further, stunts or challenges may overlap with content that encourages acts of self-harm (see Section 4) or suicide (see Section 3).

10.5 In evaluating the risk of a particular stunt or challenge, service providers should consider both the likelihood of children copying the stunt or challenge, and the likelihood that it could result in serious injury to children (even if judged to be safe for adults).

10.6 Consistent with freedom of expression, there are many types of content containing stunts and challenges that should not be considered to be dangerous or harmful to children

91 Definition of ‘stunt’, adapted from Collins Dictionary, n.d. [accessed 15 April 2025]. 92 Definition of ‘challenge’, adapted from Oxford English Dictionary, n.d. [accessed 15 April 2025].

because of relevant contextual factors. For example, stunts and challenges undertaken by highly-skilled or professional individuals, where these are not instructive or encouraged, are unlikely to be considered dangerous stunts and challenges content. Content may depict challenges (including those to raise money for charitable reasons) which involve no risk of serious injury or have been undertaken by adults who have trained and/or understood the risks involved: such content is also unlikely to be considered dangerous stunts and challenges content.

10.7 Certain contextual factors can be used to determine whether a challenge is likely to be emulated by children. Participation in challenges and stunts is often driven by social pressure, and content which involves direct encouragement by peers to take part is more likely to encourage children to emulate the challenge or stunt.

10.8 The information or commentary presented within or alongside the stunt or challenge content can impact whether a child is likely to emulate the stunt or challenge. Service providers should also consider factors which might make challenges easier to emulate, such as whether the content provides instructions for the dangerous stunt or challenge which are sufficiently detailed. An absence of important safety-related instructions or information relating to a dangerous stunts or challenges may also increase the likelihood of serious injury.

10.9 Some content relating to challenges or stunts may include commentary seeking to reduce the risk of harm. For example, services may encounter videos depicting a dangerous challenge or stunt, which contain warnings discouraging emulation (e.g., ‘do not try this at home’). These content warnings may be relevant as contextual factors. However, services should note that such warnings can be used to avoid detection and/or attempt to limit the personal liability of the individual posting. Warnings may not be sufficient to mitigate the risk of harm if the content includes other features that encourage children to enact stunts and challenges likely to result in serious injury.

Examples¶

Examples or kinds of content that Ofcom considers to be dangerous stunts and challenges content that is harmful to children¶

10.10 Table 10.2 describes the kinds of dangerous stunts and challenges content Ofcom considers to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

10.11 We have not addressed the level or impact of harm within Table 10.2. For further information on the risks and impact of harm, please see Section 9 of the Children’s Register of Risks.

10.12 Many kinds of content will only meet the definition under the Act in certain, limited circumstances, which we illustrate with examples in the table.

Table 10.2: Descriptions and examples of dangerous stunts and challenges content that is harmful to children

Content description

Content examples

Encouraging, or providing instructions for, others to partake in a challenge that could cause serious injury

Content encouraging, or providing instructions for, others

to take part in challenges such as the following:

• Asphyxiation challenges (e.g., blackout challenge,

choking challenge).

• Ingestion challenges (e.g., drinking large

quantities of alcohol or liquid medication, drinking

or eating harmful substances).

• Challenges that encourage individuals to apply

harmful substances to their skin where this could

result in serious injury.

• Challenges to unsafely climb tall buildings or

sit/stand on the edge of tall buildings and

structures.

Encouraging others to emulate stunts carried out by professionals that could cause serious injury if emulated

• Content encouraging untrained individuals to replicate stunts undertaken by professionals in films or for entertainment, e.g., jumping out of a moving vehicle. • Content encouraging individuals to replicate dangerous stunts in the context of extreme sports (e.g., snow sports, climbing, skateboarding or parkour), when the action is likely to cause serious harm if attempted by someone who has not been trained, and the post does not contain sufficient commentary relating to safety.

Content description

Content examples

Challenges which involve no risk of serious injury

• Challenges which involve no risk or serious injury, e.g.,

charitable challenges such as the ‘ALS Ice Bucket

Challenge’, where participants were encouraged to

pour ice-cold water over themselves to promote

awareness of ALS (amyotrophic lateral sclerosis).

Challenges or stunts which are not likely to be encouraged

• Depictions of stunts undertaken by a professional, where there are no instructions to carry them out, and where these are not promoted or otherwise encouraged, e.g., a clip from a film. • Content depicting stunts and challenges in the context of extreme sports that, if instructional, provides sufficient safety-related information to mitigate the risk of serious harm if enacted. • Content depicting stunts and challenges in the context of extreme sports that is not instructional and/or does not include wording that encourages others to attempt it.

Examples or kinds of content Ofcom does not consider to be dangerous stunts and challenges content that is harmful to children¶

10.13 Table 10.3 describes the kinds of dangerous stunts and challenges content that Ofcom does not consider to meet the definition of PC under the Act, along with a non-exhaustive list of examples.

10.14 Although Ofcom does not consider the examples in the following table to meet the definition of PC, the content may still be inappropriate for children in certain age groups.

Table 10.3: Descriptions and examples of content that is not dangerous stunts and challenges content that is harmful to children