In our recent analysis of Ofcom’s advice to the Secretary of State on thresholds for categorisation under the Online Safety Act, we noted that “none of the evidence submitted by civil society organisations to Ofcom in their call for evidence (published here) is referenced in their research. This includes submissions from the Antisemitism Policy Trust and Carnegie UK, both of whom were instrumental in supporting Baroness Morgan to win her concession, 5 Rights (which is headed up by Baroness Kidron) and from Refuge and Glitch, who both reference Morgan’s amendment in their responses. Industry submissions are referenced as being considered, however …”

We flagged that we would provide further analysis on the submitted evidence and how much it is reflected in the subsequent advice, which we do so below. We are grateful for the input of Harry Wild to the research and preparation of this note.

Summary

In July 2023, Ofcom issued a call for evidence to help “form the basis for Ofcom’s advice to Government”. Their call for evidence set out that they were “keen to gather evidence and information from stakeholders to help inform our research set out below. We will then use the responses to inform our research and ultimately our advice to Government on categorisation.” The regulator subsequently published the 30 responses that they had received (or at least the responses that were not badged as confidential). These responses came from various organisations including charities, retailers, social media platforms, and search services.

However, in the published advice which flowed from the 2023 call for evidence, there were only six references to evidence received in that call - all of them from industry. (See paragraphs 2.13, 3.13, 5.3 (it is “currently challenging for services to accurately identify and measure the number of child users”), A6.6, A6.10 (“we have sought (where possible) to ensure our approach accommodates existing industry practice”) and A6.15.)

As we set out in our analysis, there are a number of aspects relating to the Parliamentary intent of the Act with regard to categorisation and - specifically - the late concession by the Government to an amendment by Baroness Morgan to give Ofcom the flexibility to consider size or functionality when providing categorisation advice to the Secretary of State. We have therefore analysed the submissions provided to Ofcom to see whether those issues were evidenced in the material they received as part of their call. Our judgement is that they were.

This throws up a number of questions as to what criteria Ofcom is applying to “evidence” (which we have talked about in our response to the illegal harms consultation) as well as to the value of organisations, particularly over-stretched, under-resourced civil society organisations, preparing material to submit to such calls when the substance of that evidence is not referenced in the response.

Our analysis of the evidence provided to Ofcom

Small platforms = high risk

One key theme throughout the responses received by Ofcom was that size should not be the main determining factor when it comes to categorisation. As we noted in our analysis, this is something that should have been captured more clearly in the advice to the Secretary of State due to Baroness Nicky Morgan’s amendment to the Bill, changing the wording from “size and functionality” to “size or functionality”.

Research from civil society organisations demonstrated how smaller sites are often a higher risk due to their extremist content not being allowed on the more mainstream sites. Many of these organisations suggested that the functionality of the platform and the features that it has should play a bigger factor in categorisation than size, as this seems to be where the majority of the risk lies. See for example, 5 Rights(“Smaller platforms with fewer users and less functionality are not by definition less risky.”)

Features that impact the speed and ease of dissemination of information, particularly on user-to-user services, were highlighted in the majority of responses (see for example, Alan Turing Institute, Antisemitism Policy Trust, Barnardos, Carnegie UK, Glitch, Refuge) and in the research that was given to OFCOM. In its advice, Ofcom looked at virality but only in the context of size, as specified in the Act.

(Some) large platforms = low risk

Conversely, many of the commercial organisations who submitted responses pointed out how, despite their size, they judged that their platform had a much lower risk level compared to other smaller sites. There was also another concern brought forward by some organisations that the Act would fundamentally put their services at risk. For example, Wikimedia, argued that their service was a “public good platform” which posed little risk, and therefore should be left out of scope of category 1 (should that be determined by size alone). During the debates in Parliament in support of Baroness Morgan’s amendment, it was noted that - as well as capturing small, high-harm platforms, as was the intent - it would also “solve the Wikipedia” problem by providing Ofcom with a means *not* to bring that platform into Category 1 due to its different functionality.

Small platforms and their resourcing problems

Several responses from industry argued that their organisations were too small to effectively comply with the Act. Reddit, Wikimedia, Trustpilot, and others brought forward arguments in support of this idea, citing smaller employee headcounts or the fact that they do not rely on wide dissemination of information for revenue. Their arguments highlight concerns about burdens - should their platforms be subjected to additional duties under category 1 - appear to have borne more weight than the arguments from civil society organisations arguing that some small platforms, set up primarily to cause harm, should be brought into the regime.

Conclusion

None of the evidence cited above regarding the categorisation of services was referenced in the final advice to the Secretary of State, nor did Ofcom use this evidence to provide any detailed analysis of their considerations as to whether to use the flexibility afforded to them by the late amendment to the Act. Given the burdens on civil society organisations - and the continued emphasis that Ofcom is placing on needing “more evidence” to address gaps or concerns flagged in its subsequent OSA consultation - this lack of consideration is disappointing.