The Online Safety Act grumbles on

Policymakers sometimes comfort themselves that if no-one is completely satisfied, they have probably got it about right. 

On that basis, Ofcom’s implementation of the Online Safety Act’s illegality duties must be near-perfection: the Secretary of State (DSIT) administering a sharp nudge with his draft Statement of Strategic Priorities, while simultaneously under fire for accepting Ofcom’s advice on categorisation of services; volunteer-led community forums threatening to close down in the face of perceived compliance burdens; and many of the Act’s cheerleaders complaining that Ofcom’s implementation has so far served up less substantial fare than they envisaged. 

As of now, an estimated 25,000 UK user-to-user and search providers (plus another 75,000 around the world) are meant to be busily engaged in getting their Illegal Harms risk assessments finished by 16 March. 

Today is Safer Internet Day. So perhaps spare a thought for those who are getting to grips with core and enhanced inputs, puzzling over what amounts to a ‘significant’ number of users, learning that a few risk factors may constitute ‘many’ (footnote 74 to Ofcom’s General Risk Level Table), or wondering whether their service can be ‘low risk’ if they allow users to post hyperlinks.  (Ofcom has determined that hyperlinks are a risk factor for six of the 17 kinds of priority offence designated by the Act: terrorism, CSEA, fraud and financial services, drugs and psychoactive substances, encouraging or assisting suicide and foreign interference offences). 

Grumbles from whichever quarter will come as no great surprise to those (this author included) who have argued from the start that the legislation is an ill-conceived, unworkable mess which was always destined to end in tears. Even so, and making due allowance for the well-nigh impossible task with which Ofcom has been landed, there is an abiding impression that Ofcom’s efforts to flesh out the service provider duties - risk assessment in particular – could have been made easier to understand. 

The original illegal harms consultation drew flak for its sheer bulk: a tad over 1,700 pages. The final round of illegal harms documents is even weightier: over 2,400 pages in all. It is in two parts. The first is a Statement. In accordance with Ofcom’s standing consultation principles, it aims to explain what Ofcom is going to do and why, showing how respondents’ views helped to shape Ofcom’s decisions. That amounts to 1,175 pages, including two summaries. 

The remaining 1,248 pages consist of statutory documents: those that the Act itself requires Ofcom to produce. These are a Register of Risks, Risk Assessment Guidance, Risk Profiles, Record Keeping and Review Guidance, a User to User Illegal Content Code of Practice, a Search Service Illegal Content Code of Practice, Illegal Content Judgements Guidance, Enforcement Guidance, and Guidance on Content Communicated Publicly and Privately. Drafts of the two Codes of Practice were laid before Parliament on 16 December 2024. Ofcom can issue them in final form upon completion of that procedure.

When it comes to ease of understanding, it is tempting to go on at length about the terminological tangles to be found in the documents, particularly around ‘harm’, ‘illegal harm’ and ‘kinds of illegal harm’. But really, what more is worth saying? Ofcom’s documents are, to all intents and purposes, set in stone. Does it help anyone to pen another few thousand words bemoaning opaque language? Other than in giving comfort that they are not alone to those struggling to understand the documents, probably not. Everyone has to get on and make the best of it.

So one illustration will have to suffice. ‘Illegal harm’ is not a term defined or used in the Act. In the original consultation documents Ofcom’s use of ‘illegal harm’ veered back and forth between the underlying offence, the harm caused by an offence, and a general catch-all for the illegality duties; often leaving the reader to guess in which sense it was being used. 

The final documents are improved in some places, but introduce new conundrums in others. One of the most striking examples is paragraph 2.35 and Table 6 of the Risk Assessment Guidance (emphasis added to all quotations below). 

Paragraph 2.35 says: 

“When evaluating the likelihood of a kind of illegal content occurring on your service and the chance of your service being used to commit or facilitate an offence, you should ask yourself the questions set out in Table 6.”

Table 6 is headed: 

“What to consider when assessing the likelihood of illegal content”

The table then switches from ‘illegal content’ to ‘illegal harm’. The first suggested question in the table is whether risk factors indicate that: 

“this kind of illegal harm is likely to occur on your service?” 

‘Illegal harm’ is footnoted with a reference to a definition in the Introduction: 

“the physical or psychological harm which can occur from a user encountering any kind of illegal content…”. 

So what is the reader supposed to be evaluating: the likelihood of occurrence of illegal content, or the likelihood of physical or psychological harm arising from such content? 

If ‘Illegal Harm’ had been nothing more than a title that Ofcom gave to its illegality workstream, then what the term actually meant might not have mattered very much. But the various duties that the Act places on service providers, and even Ofcom’s own duties, rest on carefully crafted distinctions between illegal content, underlying criminal offences and harm (meaning physical or psychological harm) arising from such illegality. 

That can be seen in this visualisation. It illustrates the U2U service provider illegality duties - both risk assessment and substantive - together with the Ofcom duty to prepare an illegality Risks Register and Risk Profiles.  The visualisation divides the duties into four zones (A, B, C and D), explained below. 

A: The duties in this zone require U2U providers to assess certain risks related to illegal content (priority and non-priority). These risks are independent of and unrelated to harm. The risks to be assessed have no direct counterpart in any of the substantive safety duties in Section 10. Their relevance to those safety duties probably lies in the proportionality assessment of measures to fulfil the Section 10 duties. 

Although the service provider’s risk assessment has to take account of the Ofcom Risk Profile that relates to its particular kind of service, Ofcom’s Risk Profiles are narrower in scope than the service provider risk assessment. Under the Act Ofcom’s Risks Register and Risk Profiles are limited to the risk of harm (meaning physical or psychological harm) to individuals in the UK presented by illegal content present on U2U services and by the use of such services for the commission or facilitation of priority offences. 

B:  This zone contains harm-related duties (identified in yellow): Ofcom Risk Profiles, several service provider risk assessment duties framed by reference to harm, plus the one substantive Section 10 duty framed by reference to harm (fed by the results of the harm-related risk assessment duties). Harm has its standard meaning in the Act: physical or psychological harm. 

C: This zone contains two service provider risk assessment duties which are independent of and unrelated to risk of harm, but which feed directly into a corresponding substantive Section 10 duty. 

D: This zone contains the substantive Section 10 duties: one based on harm and three which stand alone. Those three are not directly coupled to the service provider’s risk assessment.

This web of duties is undeniably complex. One can sympathise with the challenge of rendering it into a practical and readily understandable risk assessment process capable of feeding the substantive duties.  Nevertheless, a plainer and more consistently applied approach to terminology in Ofcom's documents would have paid dividends.

Safe speech by design

Proponents of a duty of care for online platforms have long dwelt on the theme of safety by design. It has come to the fore again recently with the government’s publication of a draft Statement of Strategic Priorities (SSP) for Ofcom under the Online Safety Act.  Safety by Design is named as one of five key areas.

Ofcom is required to have regard to the final version of the SSP in carrying out its functions under the Act. Given Ofcom’s regulatory independence the government can go only so far in suggesting how Ofcom should do its job. But, in effect, the SSP gives Ofcom a heavy hint about the various directions in which the government would like it to go.

So what does safety by design really mean? How might it fit in with platform (U2U) and search engine duties under the Online Safety Act (OSA)?

Before delving into this, it is worth emphasising that although formulations of online platform safety by design can range very widely [1] [2], for the purposes of the OSA safety by design has to be viewed through the lens of the specific duties imposed by the Act.

This piece focuses on the Act’s U2U illegality duties. Three of the substantive duties concern design or operation of the service:

• Preventing users encountering priority illegal content by means of the service (S. 10(2)(a))
• Mitigating and managing the risk of the service being used for the commission or facilitation of a priority offence (as identified in the most recent illegal content risk assessment of the service) (S. 10(2)(b))
• Mitigating and managing the risks of physical or psychological harm to individuals (again as identified in the most recent illegal content risk assessment) (S. 10(2)(c))

Two further substantive illegality duties are operational, relating to:

• Minimising the length of time for which priority illegal content is present on the service (S. 10(3)(a))
• Swiftly taking down illegal content where the service provider is alerted to or otherwise becomes aware of its presence. (S. 10(3)(b))

S.10(4) of the Act provides examples of the areas of design, operation and use of a service to which the duties apply and, if proportionate, require the service provider to take or use measures. Those include “design of functionalities, algorithms and other features.”

Safety by design in the Online Safety Act

When applied to online speech, the notion of safety by design prompts some immediate questions: What is safety? What is harm?

The OSA is less than helpful about this. It does not define safety, or safety by design. It defines harm as physical or psychological harm, but that term appears in only one of the five substantive illegality duties outlined above. Harm has more a pronounced, but not exclusive, place in the prior illegal content risk assessment that a platform is required to undertake.

Safety by design gained particular prominence with a last-minute House of Lords addition to the Bill: an introductory ‘purpose’ clause. This amendment was the result of cross-party collaboration between the then Conservative government and the Labour Opposition.

What is now Section 1 proclaims (among other things) that the Act provides for a new regulatory framework which has the:

“general purpose of making the use of internet services regulated by this Act safer for individuals in the United Kingdom.”

It goes on that to achieve that purpose, the Act (among other things):

“imposes duties which, in broad terms, require providers of services regulated by this Act to identify, mitigate and manage the risks of harm (including risks which particularly affect individuals with a certain characteristic) from

(i) illegal content and activity, and

(ii) content and activity that is harmful to children, …”

Finally, and most relevantly, it adds that:

“Duties imposed on providers by this Act seek to secure (among other things) that services regulated by this Act are … safe by design…”.

A purpose clause is intended to assist in the interpretation of the legislation by setting out the purposes for Parliament intended to legislate, rather than leaving the courts to infer them from the statutory language.

Whether such clauses in fact tend to help or hinder is a matter of lawyerly debate. This clause is especially confusing in its use of terms that are not defined by the Act and do not have a clear and obvious ordinary meaning (“safe” and “safe by design”), mixed up with terms that are specifically defined in the legislation (“harm”, meaning physical or psychological harm).

One thought might be that “safe” means safe from physical or psychological harm, and that “safe by design” should be understood accordingly. However, that seems unlikely since four of the five substantive illegality duties on service providers relate to illegal content and activity per se, irrespective of whether they might involve a risk of physical or psychological harm to individuals.

S.235 defines Ofcom’s “online safety functions” in terms of all its functions under the Act. In contrast, the transitional provisions for Video Service Providers define “safety duties” in terms focused on platform duties in respect of illegality and harm to children.

Similarly, in the earlier part of the Act only those two sets of duties are described (albeit merely in the section headings) as “safety duties”. “Safe by design” may possibly refer to those duties alone.   

The concept of safety by design tends to embody some or all of a number of elements: risk-creating features; prevention and reduction of harm; achieving those by appropriate design of a risk-creating feature, or by adding technical safeguards.

The most general aspect of safety by design concerns timing: that safety should be designed in from the outset rather than thought about afterwards.

Prevention itself has a temporal aspect, but that may relate as much to the kind of measure as to the stage of development at which it should be considered. Thus the Minister’s introduction to the Statement of Strategic Priorities says that it:

“includes ensuring safety is baked into platforms from the start so more harm is caught before it occurs”.

This could refer to the point at which a safety measure intervenes in the user’s activity, as opposed to (or as well as) the stage at which the designers consider it.

Later in the Statement, safety by design is expressly said to include deploying technology in content moderation processes. Providers would be expected to:

“…embed proportionate safety by design principles to mitigate the [risk of their service being used to facilitate illegal activity]. This should include steps such as … where proportionate, deploying technology to improve the scale and effectiveness of content moderation, considering factors including providers’ capacity and users’ freedom of expression and privacy rights.”

An analogy with product safety could suggest that safety by design is about identifying risk-creating features at the design stage and either designing those features in the safest way or incorporating safeguards. That aspect is emphasised by Professor Lorna Woods in a recent paper [3]:

“The objective of ‘safety by design’ is – like product safety – to reduce the tendency of a given feature or service to create or exacerbate such issues.”

Applied to products like cars that would mean that you should consider at the outset where safely to position the fuel tank, not unthinkingly place it somewhere dangerous and try to remedy the problem down the line, or after an accident has happened. Or, if a piece of machinery has a sharp cutting blade, consider at the outset how to add a guard into the design. A culture of safety by design should help to ensure that potential safety risks are considered and not overlooked.  

However, a focus on risk-creating features gives rise to particular difficulties when safety by design is translated to online platforms.

The underlying duty of care reasons for this have been rehearsed on previous occasions (here and here). In short, speech is not a tripping hazard, nor is it a piece of machinery. A cutting machine that presents a risk of physical injury to its operator is nothing like a space in which independent, sentient human beings can converse with each other and choose what to say and do.

Professor Woods [3] suggests that ‘by design’ seeks to ensure that products respect the law (my emphasis). If that is right, then by the same token it could be said that safety by design when applied to online platforms seeks to ensure that in their communications with each other users respect the law (or boundaries of harm set by the legislation). That is a materially different exercise, for which analogies with product safety can be taken only so far.

The June 2021 DCMS/DSIT paper Principles of safer online platform design opened with the statement that:

“Online harms can happen when features and functions on an online platform create a risk to users’ safety.”

For the illegality duties imposed by the OSA, when we set about identifying concrete features and functionalities that are said to create or increase risk of illegality, we run into problems when we move beyond positive platform conduct such as recommender and curation algorithms.

The example of recommender and curation algorithms has the merit of focusing on a feature that the provider has designed and which can causally affect which user content is provided to other users.

But the OSA duties of care – and thus safety by design - go well beyond algorithmic social media curation, extending to (for instance) platforms that do no more than enable users to post to a plain vanilla discussion forum.

Consider the OSA safety duties concerning priority illegal content and priority offences.  What kind of feature would create or increase a risk of, for example, an online user deciding to offer boat trips across the Channel to aspiring illegal immigrants?

The further we move away from positive content-related functionality, the more difficult it becomes to envisage how safety by design grounded in the notion of specific risk-creating features and functions might map on to real-world technical features of online platforms.

The draft SSP confirms that under the OSA safety by design is intended to be about more than algorithms:

“When we discuss safety by design, we mean that regulated providers should look at all areas of their services and business models, including algorithms and functionalities, when considering how to protect all users online. They should focus not only on managing risks but embedding safety outcomes throughout the design and development of new features and functionalities, and consider how to make existing features safer.”

Ofcom faced the question of risk-creating features when preparing the risk profiles that the Act requires it to provide for different kinds of in-scope service. For the U2U illegality risk profile it has to:

“carry out risk assessments to identify and assess the following risks of harm presented by [user to user] services of different kinds—

(a) the risks of harm to individuals in the United Kingdom presented by illegal content present on regulated user-to-user services and by the use of such services for the commission or facilitation of priority offences; …”

The risks that Ofcom has to identify and assess, it should be noted, are not the bare risk of illegal content or illegal activity, but the risk of harm (meaning physical or psychological harm) to individuals presented by such content or activity.

Ofcom is required to identify characteristics of different kinds of services that are relevant to those risks of harm, and to assess the impact of those kinds of characteristics on such risks. “Characteristics” of a service include its functionalities, user base, business model, governance and other systems and processes.

Although a platform has to carry out its own illegal content risk assessment, taking account of Ofcom’s risk profile, the illegality risks that the platform has to assess also include bare (non-harm-related) illegality.

Ofcom recognises that functionalities are not necessarily risk-creating:

“Functionalities in general are not inherently positive nor negative. They facilitate communication at scale and reduce frictions in user-to-user interactions, making it possible to disseminate both positive and harmful content. For example, users can engage with one another through direct messaging and livestreaming, develop relationships and reduce social isolation. In contrast, functionalities can also enable the sharing of illegal material such as livestreams of terrorist atrocities or messages sent with the intent of grooming children.” [6W.16]

Ofcom overcomes this issue in its proposed risk profiles by going beyond characteristics that of themselves create or increase risks of illegality. This is most clearly expressed in Volume 2 of its Illegal Harms Consultation:

“We recognise that not all characteristics are inherently harmful; we therefore use the term ‘risk factor’ to describe a characteristic for which there is evidence of a risk of harm to individuals. For example, a functionality like livestreaming is not inherently risky but evidence has shown that it can be abused by perpetrators; when considering specific offences such as terrorism or CSEA, a functionality like livestreaming can give rise to risk of harm or the commission or facilitation of an offence.” [5.26]

General purpose functionality and features of online communication can thus be designated as risk factors, on the basis that there is evidence that wrongdoers make use of them or, in some instances, certain combinations of features.

Since measures focused on general purpose features are likely to be vulnerable to objections of disproportionate interference with freedom of expression, for such features the focus of preventing or mitigating the identified risk is more likely to be on other aspects of the platform’s design, on user options and controls in relation to that feature (e.g. an option to disable the feature), or on measures such as content moderation.

Ofcom implicitly recognises this in the context of livestreaming:

“6.11 We acknowledge that some of the risk factors, which the evidence has demonstrated are linked to a particular kind of illegal harm, can also be beneficial to users. This can be in terms of the communication that they facilitate, or in some cases fulfilling other objectives, such as protecting user privacy. …

6.13 While livestreaming can be a risk factor for several kinds of illegal harm as it can allow the real-time sharing of illegal content, it also allows for real-time updates in news, providing crucial information to a wide-range of individuals.

6.14 These considerations are a key part of the analysis underpinning our Codes measures.”

The result is that while the illegality risk profiles that Ofcom has proposed include as risk factors a range of platform features that could be viewed as general purpose, they tend not to translate into recommended measures aimed at inhibiting that feature.

Here is a selection of features included in the proposed illegality risk profile:

Service feature	Risk (likelihood of increased risk of harm related to offences involving):
Ability to create user profiles	Grooming, harassment, stalking, threats, abuse, drugs and psychoactive substances, unlawful immigration, human trafficking, sexual exploitation of adults; and for the risk of fake profiles: Grooming, harassment, stalking, threats, abuse, controlling or coercive behaviour, proceeds of crime, fraud and financial services, foreign interference offences.
Users can form user groups	Grooming, encouraging or assisting suicide or serious self-harm, drugs and psychoactive substances, unlawful immigration, human trafficking.
Livestreaming	Terrorism, grooming, image-based CSAM, encouraging or assisting suicide or serious self-harm, harassment, stalking, threats, abuse.
Direct messaging	Grooming and CSAM, hate, harassment, stalking, threats, abuse, controlling or coercive behaviour, intimate image abuse, fraud and financial services offences.
Encrypted messaging	Terrorism, grooming, CSAM, drugs and psychoactive substances, sexual exploitation of adults, foreign interference, fraud and financial services offences.
Ability to comment on content	Terrorism, grooming, encouraging or assisting suicide or serious self-harm, hate, harassment, stalking, threats, abuse.
Ability to post images or videos	Terrorism, image-based CSAM, encouraging or assisting suicide or serious self-harm, controlling or coercive behaviour, drugs and psychoactive substances, extreme pornography, intimate image abuse.
Ability to repost or forward content	Encouraging or assisting suicide or serious self-harm, harassment, stalking, threats, abuse, intimate image abuse, foreign interference.
Ability to search for user generated content	Drugs and psychoactive substances, firearms and other weapons, extreme pornography, fraud and financial services offences.
Hyperlinks	Terrorism, CSAM URLs, foreign interference offences.

Other functionality risk factors include anonymity, user connections (such as friending and following), group messaging, and ability to post or send location information.

Designation of general purpose functionality as a risk factor reaches a high point with hyperlinks. Since terrorists and other potential perpetrators can use hyperlinks to point people to illegal material, hyperlinks can be designated as a risk factor despite not being inherently harmful.

It is worth recalling what the ECtHR said in Magyar Jeti Zrt (ECtHR) about the central role of hyperlinks in internet communication:

“Furthermore, bearing in mind the role of the Internet in enhancing the public’s access to news and information, the Court points out that the very purpose of hyperlinks is, by directing to other pages and web resources, to allow Internet users to navigate to and from material in a network characterised by the availability of an immense amount of information. Hyperlinks contribute to the smooth operation of the Internet by making information accessible through linking it to each other.”

General purpose functionality as a risk factor was foreshadowed in the June 2021 DCMS paper. Arguably it went further, asserting in effect that providing a platform for users to communicate with each other is itself a risk-creating activity:

          “Your users may be at increased risk of online harms if your platform allows them to:

• interact with each other, such as through chat, comments, liking or tagging
• create and share text, images, audio or video (user-generated content)”

In the context of the internet in the 21st century, this list of features describes commonplace aspects of the ability to communicate electronically. In a former age we might equally have said that pen, paper, typewriter and the printing press are risk factors, since perpetrators of wrongdoing may use written communications for their nefarious purposes.

Whilst Ofcom recognises the potential freedom of expression implications of treating general purpose functionalities as illegality risk factors, it always has to be borne in mind that from a fundamental rights perspective the starting point is that speech is a right, not a risk. Indeed the Indian Supreme Court has held that the right of freedom of expression includes the reach of online individual speech:

"There is no dispute that freedom of speech and expression includes the right to disseminate information to as wide a section of the population as is possible."

That is not to suggest that freedom of expression is an absolute right. But any interference has to constitute a sufficiently clear and precise rule (especially from the perspective of the user whose expression is liable to be interfered with), then satisfy necessity and proportionality tests.

Preventative technological measures

A preventative approach to safety by design can easily lean towards technological measures: since this is a technology product, technological preventative measures should be designed in to the service and considered at the outset.

Professor Woods [3], argues that:

“Designing for safety (or some other societal value) does not equate to techno-solutionism (or techno-optimism); the reliance on a “magic box” to solve society’s woes or provide a quick fix.”

However, in the hands of government and regulators it has a strong tendency to do so.[4].  Indeed the draft SSP devotes one of its five key priorities to Technology and Innovation, opening with:

“Technology is vital to protecting users online and for platforms fulfilling their duties under the Act.”

Later:

“It is not enough that new, innovative solutions to known problems exist – online service providers must also adopt and deploy these solutions to improve user safety. … The government … encourages Ofcom to be ambitious in its [code of practice] recommendations and ensure they maintain pace with technology as it develops.”

We have already seen that in the draft SSP, safety by design is said to include deploying technology in content moderation processes.

On the basis of prevention, an inbuilt technological design measure that reduces the amount of (or exposure to) illegal user speech or activity should be preferable to hiring legions of content moderators when the platform starts operating.

However, translating duties of care or safety by design into automated or technology-assisted content moderation can come into conflict with an approach in which non-content-specific safety features are seen as preferable.

Professor Woods said in the same paper:

“At the moment, content moderation seems to be in tension with the design features that are influencing the creation of content in the first place, making moderation a harder job. So, a “by design” approach is a necessary precondition for ensuring that other ex post responses have a chance of success.

While a “by design” approach is important, it is not sufficient on its own; there will be a need to keep reviewing design choices and updating them, as well as perhaps considering ex post measures to deal with residual issues that cannot be designed out, even if the incidence of such issues has been reduced.”

As to what ex post measures might consist of, in a letter to The Times in August, Professor Woods said:

“Through a duty of care, service operators are required to ensure that their products are as safe as reasonably possible and to take steps to mitigate unintended consequences. Essentially this is product safety, or health and safety at work. This approach allows a range of interventions that do not rely on content take-down and, indeed, could be content-neutral. One example might be creator reward programmes that incentivise the spreading of clickbait material. (emphasis added)].

Maeve Walsh, writing for the Online Safety Network shortly before publication of the draft SSP [5], contrasted safety by design with thinking about the OSA “primarily as a takedown-focused regime, centering on individual pieces of content.”

Content-neutrality suggests that a safety measure in relation to a functional feature should, rather than relating specifically to some kind of illegal or harmful content, either have no effect on content as such or, if it does affect user content, do so agnostically.

Some measures have no direct effect on user content: a help button would be an example. Others may affect content, but are not targeted at particular kinds of content: for instance, a friction-reducing measure like capping the permissible number of reposts, or other measures inhibiting virality.

A measure such as a quantititive cap on the use of some feature has the advantage from a rule of law perspective that it can be clearly and precisely articulated. However, by virtue of the fact that it constrains legitimate as well as illegitimate user speech across the board, it is potentially vulnerable to proportionality objections.

Thanks to the difficulty of making accurate illegality judgements, automated content filtering and blocking technologies are potentially at risk on both scores.

[1] Trust & Safety Professional Association. Safety by Design Curriculum chapter.

[2] Australian eSafety Commissioner. Safety by Design.

[3] Professor Lorna Woods, for the Online Safety Network (October 2024). Safety by Design. 

[4] Maria P. Angel, danah boyd (12 March 2024). Proceedings of 3rd ACM Computer Science and Law Symposium (CSLAW’24) Techno-legal Solutionism: Regulating Children’s Online Safety in the United States.

[5] Maeve Walsh, for the Online Safety Network (11 October 2024). Safety by design: has its time finally come? 

Data Protection meets the Online Safety Act

This is sixth and final instalment in a series of reflections on Ofcom’s Illegal Harms consultation under the Online Safety Act 2023. Ofcom is due to publish the final version of its Illegal Harms Codes of Practice and Guidance in December.

The interaction between data protection law and the Online Safety Act’s illegal content duties attracted almost no attention during the passage of the Bill through Parliament. Nor does data protection garner more than a bare mention in the body of the Act itself. Nevertheless, service providers will have to perform their obligations compatibly with data protection laws.

However, data protection law does not sit entirely neatly alongside the OSA. It overlaps and potentially collides with some of the substantive measures that the Act requires service providers to take. This creates tensions between the two regimes.

As the process of implementing the OSA’s service provider duties has got under way, more attention has been directed to how the two regimes fit together.

At the most general level, while the Bill was still under discussion, on 25 November 2022 the ICO and Ofcom published a Joint Statement on online safety and data protection. This was an aspirational document, setting out shared goals of maximising coherence between the data protection and online safety regimes and working together to promote compliance with them. It envisaged a renewed formal memorandum of understanding between the ICO and Ofcom (yet to appear).  A more detailed Joint Statement on collaboration between the two regulators was issued on 1 May 2024.

The November 2022 statement recognised that:

“there are sometimes tensions between safety and privacy. For example, to protect users’ safety online, services might need to collect more information about their users, the content they view and their behaviour online. To protect users’ privacy, services can and should limit this data collection to what is proportionate and necessary.”

It went on:

“Where there are tensions between privacy and safety objectives, we will provide clarity on how compliance can be achieved with both regimes.”

On 16 February 2024, a week before the end of Ofcom’s Illegal Harms consultation period, the Information Commissioner’s Office published 47 pages of guidance on how data protection law applies to online content moderation processes - including moderation carried out to comply with duties under the Online Safety Act. It avowed an aim to support organisations carrying out content moderation in scope of the Act. Four months later, the ICO invited feedback on its Guidance.

Ofcom’s Illegal Harms consultation is itself liberally garnished with warnings that data protection law must be complied with, but less generously endowed with concrete guidance on exactly how to do so.

The ICO Guidance, although it put some flesh on the bones, was still pitched at a relatively high level. That was a deliberate decision. The accompanying Impact Assessment records that the ICO considered, but rejected, the option of:

“more extensive guidance discussing in depth how data protection law applies when developing or using content moderation”

in favour of:

“High level guidance setting out the ICO’s preliminary data protection and privacy expectations for online content moderation, and providing practical examples, with plans for further work as the policy area develops”.

The reason for this decision was that:

“it provides some degree of clarity for a wide variety of stakeholders, whilst still allowing the necessary flexibility for our policy positions to develop during the early stages of Ofcom’s policy and guidance development”.

The next, and perhaps most interesting, document was the ICO’s own submission to Ofcom’s Illegal Harms consultation, published on 1 March 2024. In this document the tensions between the OSA and data protection are most evident. In some areas the ICO overtly took issue with Ofcom’s approach.

What are some of the potential areas of tension?

Illegality risk assessment

The Ofcom consultation suggests that for the illegality risk assessment required under S.9 OSA service providers should, among other things, consider the following ‘core input’ to the risk assessment:

“assess any other evidence they hold that is relevant to harms on their service. This could include any existing harms reporting, research held by the service, referrals to law enforcement, data on or analysis of user behaviour relating to harms or product testing. Any types of evidence listed under Ofcom’s enhanced inputs (e.g. the results of content moderation, product testing, commissioned research) that the business already collects and which are relevant to the risk assessment, should inform the assessment. In effect, if the service already holds these inputs, they should be considered as core inputs.” [Table 9.4]

Ofcom adds that:

“… any use of users’ personal data (including any data that is not anonymised), will require services to comply with their obligations under UK data protection law. Services will need to make judgments on the data they hold to ensure it is processed lawfully, including providing appropriate transparency to users when the data is collected or further processed.” [Table 9.4]

The topic of core inputs caught the eye of the ICO, which observed:

“A key data protection consideration when processing personal data for risk assessment is the data minimisation principle set out in Article 5(1)(c) of the UK GDPR. This requires the personal data that services process to be adequate, relevant and limited to what is necessary in relation to the purposes for which it is processed. This means that services should identify the minimum amount of personal data they need to fulfil their purpose.” [p.7 – p.8]

Illegality judgements and data minimisation

Data minimisation is, more generally, an area of potential tension between the two regimes.

The Act requires service providers to make judgements about the illegality of user content. The less information is available to a service provider, the greater the risk of making arbitrary judgements (with potential ECHR implications). But the more information is retained or collected in order to make better judgements, the greater the potential conflict with the data minimisation principle (UK GDPR Article 5(1)(c)).

Section 192 of the OSA requires service providers to make illegality judgments on the basis of all relevant information reasonably available to a service provider. Ofcom’s Illegal Harms consultation document suggests that what is regarded as “reasonably available” may be limited by the constraints of data protection law:

“However, we recognise that in certain instances services may have access to information, which is relevant to a specific content judgement, but which is not either typically available to all services, which would require significant resources to collect, or the use of which would not be lawful under data protection or privacy laws. When making illegal content judgments, services should continue to have reasonable regard to any other relevant information to which they have access, above and beyond what is set out in the Illegal Content Judgements Guidance but only so long as this information is processed lawfully, including in particular in line with data protection laws.” [26.27]

The ICO cited this (and a related more equivocal passage at [A1.67]) as an example of where the Ofcom guidance is “less clear about the approach that services should take to balancing the need to make [illegal content judgements] with the need to comply with data minimisation.” The ICO said:

“The data minimisation principle requires that personal data being processed be relevant, adequate, and limited to what is necessary. Where an [illegal content judgement] can be made accurately without the need to process the additional personal data held by a service it would not be necessary for a service to process this information under data protection law. …

The text could also clarify that services may not always need to consult all available information in every instance, if it is possible to make an accurate judgement using less information.” [page 25)

There is, however, a lurking paradox of unknown unknowns. For an offence of a kind for which factual context is important, the service provider knows that relevant contextual information could exist, but does not know if it does exist. Such information (if it does exist) may or may not be available to the service provider: it could be wholly off platform and beyond its knowledge; or it could be accessible on the platform in principle, but potentially constrained by data protection law.

Without knowing whether further relevant contextual information in fact exists, how is a provider to determine whether it is able to make an accurate judgement with only the information that it knows about? How can a provider know whether further relevant information exists without going looking for it, potentially breaching data protection law in the process? The ICO Guidance says:

“You are complying with the data protection minimisation principle, as long as you can demonstrate that using [other contextual] information is: - necessary to achieve your purpose (e.g. because it ensures your decisions are accurate and fair ..."

Automated content moderation

The OSA contemplates that an Ofcom Code of Practice may for some use cases recommend service providers to undertake fully automated content moderation. The UK GDPR contains specific provisions in Article 22 about certain solely automated processing of personal data.

The Ofcom consultation says:

“Insofar as services use automated processing in content moderation, we consider that any interference with users’ rights to privacy under Article 8 ECHR would be slight. Such processing would need to be undertaken in compliance with relevant data protection legislation (including, so far as the UK GDPR applies, rules about processing by third parties or international data transfers).” [12.72]

Similar statements are made elsewhere in the consultation.

The ICO disagrees with the first sentence:

“From a data protection perspective, we do not agree that the potential privacy impact of automated scanning is slight. Whilst it is true that automation may be a useful privacy safeguard, the moderation of content using automated means will still have data protection implications for service users whose content is being scanned. Automation itself carries risks to the rights and freedoms of individuals, which can be exacerbated when the processing is carried out at scale.” [p.12]

It goes on:

“Our guidance on content moderation is clear that content moderation involves personal data processing at all stages of the moderation process, and hence data protection must be considered at all stages (including when automated processing is used, not just when a human looks at content). [p.12]

The ICO took the view that from a data protection law perspective, Ofcom’s proposed safeguards for the three recommended automated content moderation measures (CSAM perceptual hash matching, CSAM URL matching and fraud fuzzy keyword detection) are incomplete.

As to UK GDPR Article 22, the ICO commented in relation to the series of recommended measures that involve (or may involve) automated processing:

“The automated content moderation measures have the potential to engage UK GDPR Article 22, particularly measures 4G [Hash matching for CSAM] and I [Keyword detection regarding articles for use in frauds]. Article 22 of the UK GDPR places restrictions about when services can carry out solely automated decision-making based on personal information where the decision has legal or similarly significant effects. … To achieve coherence across the regimes it is important that the recommended code measures are compatible with UK GDPR Article 22 requirements.”

It is worth noting that UK GDPR Article 22 permits such solely automated decisions to be made where required by “domestic law”, provided that this sets out suitable safeguards. The new Data (Use and Access) Bill includes some changes to the Article 22 provisions.

Accuracy of illegal content judgements

Data protection law also encompasses an accuracy principle – UK GDPR Article 5(1)(d). The ICO Guidance assesses the application of this principle to be limited to the accuracy of facts underlying content judgements and to accurate recording of those judgements. However, the ICO Guidance also appears to suggest that the separate fair processing principle (UK GDPR Article 5(1)(a)) could have implications for the substantive accuracy of judgements themselves:

“You are unlikely to be treating users fairly if you make inaccurate judgements or biased moderation decisions based on their personal information."

If substantive accuracy could be both a fair processing matter and an OSA issue, how might this manifest itself? Some examples are:

Reasonably available information As already noted, S.192 of the OSA requires a service provider to make illegality judgements on the basis of all information reasonably available to it. Again as already noted, what is regarded as ‘reasonably available’ may be affected by data protection law, especially the data minimisation requirement to demonstrate that processing the data is necessary to achieve the purpose of an accurate and fair illegality judgement.

As to necessity, could the ‘unknown unknowns’ paradox (see above) come into play: how can a provider know if contextual information is available and relevant to the accuracy of the judgement that it has to make without seeking out the information? Could necessity be established simply on the basis that it is the kind of offence to which contextual information (if it existed) could be relevant, or would there have to be some justification specific to the facts of the matter under consideration, such as an indication that further information existed?

Generally, in connection with the information that service providers may use to make illegal content judgements, Ofcom says:

“When making illegal content judgements, services should continue to have reasonable regard to any other relevant information to which they have access, above and beyond what is set out in the Illegal Content Judgements Guidance but only so long as this information is processed lawfully, including in particular with data protection laws”; [26.27]

One example of this area of potential tension between the OSA and data protection is reference back to previous user complaints when making a judgement about content. In its section on handling user complaints, Ofcom’s consultation says:

“To the extent that a service needed to retain information to process complaints, this may include personal data. However, we are not proposing to recommend that services should process or retain any extra information beyond the minimum needed to comply with duties which are clearly set out on the face of the Act. To the extent that services choose to do so, this data would be held by the service subject to data protection laws.” [16.113]

Elsewhere, Ofcom says that “depending on the context, reasonably available information may include … complaints information” ([26.26, A1.66], subject to the proviso that:

“processing some of the types of information (‘data’) listed below has potential implications for users’ right to privacy. Services also need to ensure they process personal data in line with data protection laws. In particular, the likelihood is high that in considering U2U content a service will come across personal data including special category data and possibly criminal offence data, to which UK data protection laws apply.” [26.25]  

The ICO Guidance said:

“Data minimisation still applies when services use personal information to make illegal content judgements under section 192 of the OSA. Under data protection law, this means you must use personal information that is proportionate and limited to what is necessary to make illegal content judgements. …

Moderation of content can be highly contextual. Sometimes, you may need to use other types of personal information (beyond just the content) to decide whether you need to take moderation action, including users’ … records of previous content policy violations. …

You are complying with the data minimisation principle, as long as you can demonstrate that using this information is:

-        necessary to achieve your purpose (eg because it ensures your decisions are accurate and fair);

-        and no less intrusive option is available to achieve this.”

The ICO submission to the Ofcom consultation says:

“Paragraphs 16.26-27 of the consultation document state that Ofcom decided not to include a recommendation for services to keep complaints data to facilitate appeals as part of this measure. However, other consultation measures require or recommend the further use of complaints data, for example the risk assessment guidance, illegal content judgements guidance… . We think that it is important that the overall package of measures make clear what information Ofcom considers necessary for services to retain and use to comply with online safety obligations. This will help services to feel confident about complying with their data protection obligations.” [page 18]

Assuming that a service provider does have access to all relevant information, if substantive accuracy of an illegal content judgement were an aspect of the data protection fair processing principle, might that connote a degree of certainty that differs from the Act’s ‘reasonable grounds to infer’ standard in S.192 or the ‘awareness’ standard in Section 10(3)(b) (if they be different from each other)?

NCA reporting

Related to both automated processing and accuracy are the ICO’s submission comments on the obligation under S.66 for a service provider who becomes aware of previously unreported UK-linked CSAM to report it to the National Crime Agency.

The Ofcom consultation notes:

“Interference with users’ or other individuals’ privacy rights may also arise insofar as the option would lead to reporting to reporting bodies or other organisations in relation to CSAM detected using perceptual hash matching technology – for example, that a user was responsible for uploading content detected as CSAM to the service.” [14.80]

Ofcom goes on:

“As explained above, use of perceptual hash matching can result in cases where detected content is a false positive match for CSAM, or a match for content that is not CSAM and has been wrongly included in the hash database. These cases could result in individuals being incorrectly reported to reporting bodies or other organisations, which would represent a potentially significant intrusion into their privacy.” [14.84]

It adds:

“It is not possible to assess in detail the potential impact of incorrect reporting of users: the number of users potentially affected will depend on how services implement hash matching; while further details of the reporting requirements under the Act are to be specified by the Secretary of State in secondary legislation. However, the option includes principles and safeguards in relation to the hash database, the configuration of the technology, and the use of human moderators that are designed to help secure that the technology operates accurately. … [14.85]

In addition, reporting bodies have processes in place to triage and assess all reports received, ensuring that no action is taken in cases relating to obvious false positives. These processes are currently in place at NCMEC and will also be in place at the Designated Reporting Body in the NCA, to ensure that investigatory action is only taken in appropriate circumstances.” [14.86]

The ICO takes a less sanguine view of the consequences of reporting a false positive to the NCA:

“Ofcom refers to the principles and safeguards in the content moderation measures as being safeguards that are designed to help secure that the technology operates accurately in connection with user reports to the NCA. Accuracy is also a relevant consideration in data protection law. The accuracy principle requires that services take all reasonable steps to ensure that the personal data they process is not incorrect or misleading as to any matter of fact. Where content moderation decisions could have significant adverse impacts on individuals, services must be able to demonstrate that they have put sufficient effort into ensuring accuracy. 

We are concerned that the safeguards in measure 4G do not differentiate between the level of accuracy that is appropriate for reports to the NCA (which carries a particular risk of serious damage to the rights, freedoms and interests of a person who is incorrectly reported) and other significant but potentially less harmful actions such as content takedown. 

Our reading of measure 4G is that it could allow for the content moderation technology to be configured in such a way that recognises that false positives will be reported to the NCA. Whilst we acknowledge that it may not be possible to completely eliminate false positives being reported, we are concerned that a margin for error could be routinely “factored into” a service’s systems and processes as a matter of course.

This is unlikely to be compatible with a service taking all reasonable steps to ensure that the personal data it processes is not inaccurate.”

One point of particular interest is the ICO’s apparent distinction between the level of accuracy for content takedown and that for reporting to the NCA. Both Section 10(3)(b) (the takedown obligation) and Section 66 (as interpreted by Section 70) use the same language to trigger the respective obligations (emphasis added):

-        A duty to operate a service using proportionate systems and processes designed to where the provider is alerted by a person to the presence of any illegal content, or becomes aware of it in any other way, swiftly take down such content. (S.10(3)(b)

-        … must operate the service using systems and processes which secure (so far as possible) that the provider reports all detected and unreported CSEA content present on the service to the NCA. … CSEA content is “detected” by a provider when the provider becomes aware of the content, whether by means of the provider’s systems or processes or as a result of another person alerting the provider.  (S.66/70)

The ICO argument appears to suggest that data protection considerations should inform the construction of the sections, with the result that the same word ‘aware’ in the two OSA provisions would connote different levels of confidence in the accuracy of the information on which a decision is based.

End to end encryption

The Bill was loudly and repeatedly criticised by privacy and civil liberties campaigners for potentially threatening the ability to use end-to-end encryption on private messaging services. The provision that gave rise to this was what is now Section 121: a power for Ofcom to require, by notice to a U2U service provider, use of accredited technology to identify and swiftly take down, or prevent individuals encountering, CSEA content, whether communicated publicly or privately by means of the service. Service providers would have the option of developing or sourcing their own equivalent technology.

Under Section 125, a notice requiring the use of accredited technology is to be taken as requiring the provider to make such changes to the design or operation of the service as are necessary for the technology to be used effectively.

The concern with these provisions was that the effect of a notice could be to require a private messaging provider to cease using E2E encryption if it was incompatible with the technology required by the notice.

The S.121 power is self-standing, separate from the Act’s safety duties on service providers. It will be the subject of a separate Ofcom consultation, scheduled for December 2024.

Concomitantly, Ofcom is prevented from using its safety duty enforcement powers to require proactive technology to be used by a private communications service (S.136(6)). Broadly speaking, proactive technology is content identification technology, user profiling technology, or behaviour identification technology.

That is also reflected in the Schedule 4 restrictions on what Ofcom can recommend in a Code of Practice:

“Ofcom may not recommend in a Code of Practice the use of the technology to analyse user generated content communicated “privately”, or metadata relating to user-generated content communicated “privately” [14.14]

Thus, in effect, the Act’s safety duties cannot be interpreted so as to require a private communications service to use proactive technology. That is a matter for the self-standing S.121 power.

The Ofcom consultation states that E2E encryption is not inherently bad. It goes on to acknowledge the benefits of E2E encryption:

“The role of the new online safety regulations is not to restrict or prohibit the use of such functionalities, but rather to get services to put in place safeguards which allow users to enjoy the benefits they bring while managing risks appropriately” [Vol 2, Introduction]

Nevertheless, it also cites E2E encryption as a risk factor. For instance, end-to-end encryption poses the risk that “offenders often use end-to-end encrypted services to evade detection” [Vol 2, Introduction] and “end-to-end encryption guarantees a user’s privacy and security of messages, but makes it harder for users to moderate for illegal content.” [6.12]; and “Private messaging services with encryption are particularly risky, as they make the exchange of CSAM harder to detect.” [6C.139] “If your service allows encrypted messaging, we would expect you to consider how this functionality can be used by potential perpetrators to avoid monitoring of communications while sharing illegal content such as CSAM or conducting illegal behaviour.” [Table 14] The theme is repeated numerous times in relation to different offences.

In relation to its specific proposal for ‘hash matching’ to detect and remove known CSAM (Child Sexual Abuse Material), Ofcom says:

“Consistent with the restrictions in the Act, this proposal does not apply to private communications or end-to-end encrypted communications. We are not making any proposals that would involve breaking encryption. However, end-to-end encrypted services are still subject to all the safety duties set out in the Act and will still need to take steps  to mitigate risks of CSAM on their services” [Overview]

The ICO did not challenge Ofcom’s evidence bases for concluding that E2E encryption was a risk. However, it said:

“We are concerned that the benefits of these functionalities are not given enough emphasis in the risk assessment guidance and risk profiles (Annex 5). These are the documents that U2U services are most likely to consult on a regular basis. We consider that there is a risk that the risk assessment process may be interpreted by some services to mean that functionalities such as E2EE and anonymity/pseudonymity are so problematic from an online safety perspective that they should be minimised or avoided. If so, the risk assessment process could create a chilling effect on the deployment of functionalities that have important benefits, including keeping users safe online.”

and:

“In summary, we therefore suggest that the guidance should make it clear that the online safety regime does not restrict or prohibit the use of these functionalities and that the emphasis is on requiring safeguards to allow users to enjoy the benefits while managing risks appropriately.

Whilst ICO’s comments do not necessarily reflect tension between the OSA and data protection regimes as such, a difference of emphasis is detectable.

Generally, critics of the Act have long argued that requiring service providers to assess illegality is a recipe for arbitrary decision-making and over-removal of legal user content, likely to constitute an unjustified interference with the right of freedom of expression.  As Ofcom and the ICO attempt to get to grips with the practical realities of the duties, data protection and privacy are now joining the fray.