Our response to Ofcom's proposals on user identity verification and the filter tool
Clean Up The Internet has submitted our response to Ofcom's consultation "Giving users more choice, control and trust in their online experience", on the additional duties for Category 1 services. You can download our full response here:
Our response focuses on two duties in the Online Safety Act that we campaigned for. Section 64 requires the largest platforms to offer all adult users the option to verify their identity. Section 15(9) requires them to offer a tool that lets users filter out accounts that haven't been verified. Ofcom's consultation includes draft guidance on how platforms should meet the verification duty, and a measure in its draft Additional Duties Code of Practice setting out how the filter should work.
In our view, the proposals as they stand fall well short, and risk producing verification schemes that are meaningless, invisible, or both. Meaningless, because each platform can decide for itself what its scheme is for, with no requirement that this has anything to do with user safety, and Ofcom sets too few baseline expectations for how schemes should work. For example, nothing requires a free route to verification, and platforms are only encouraged, not required, to record how they designed their scheme and why. Invisible, because nothing requires platforms to tell users the scheme exists, to make it easy to find, or to show other users who is verified.
The proposed filter tool measure has similar problems. Ofcom says, in footnotes rather than in the Code itself, that an account verified under any scheme a platform runs can pass the filter, however weak that scheme is. Accounts that simply describe themselves as businesses or organisations would fall outside the filter altogether, with no check that the claim is true, and nothing says how automated accounts should be treated. Platforms could also leave parts of their service outside the filter where they judge it would interfere with their "core offering". That is a test Ofcom has added, which is not in the Act, which Ofcom has left to platforms to define, and which sits on an equal footing with user safety. And as with verification, nothing requires the filter to be easy to find, or for users to be told it exists.
To see what this could mean in practice, our submission explores a potential case Ofcom itself highlights: a woman on a large social media service facing a sustained pile-on, much of it from accounts she can't identify. This is far from rare: a Refuge and EVAW survey published in September found that 9% of women who had experienced violence against women and girls had been harmed online by a group of strangers, rising to 17% of Black and minoritised women.
If a woman in this situation decides she wants to hear only from accounts whose identity the service has checked, she first has to find the filter - and she may never have been told it exists. If she does find it, she may face several separate controls in different places, and nothing tells her whether she has switched them all on. The platform doesn't have to tell her what the filter covers, and may have left out the very features through which the pile-on is reaching her. In some places she may get a warning, or have messages moved to a separate inbox, rather than having the content filtered out. And even where the filter does apply, it can let through accounts "verified" under weak schemes, accounts claiming to be businesses or news outlets, bots, and verified accounts that have since been hacked.
This means she may end up only minimally less exposed than before. Worse, she may believe otherwise, and relax other precautions as a result. None of this would fall foul of Ofcom's draft rules as they stand.
We do believe all of this is fixable, and that Ofcom's current proposals, whilst flawed, do provide a foundation to build on. The draft guidance's reliability principle, its approach to inclusivity, and the precision Ofcom has brought to blocking and muting elsewhere in the same Code all show what is possible.
Our response makes 34 recommendations. Most suggest setting baseline expectations: requirements that don't depend solely on what each platform chooses for itself, while still leaving platforms plenty of flexibility in how they meet them.
The most fundamental recommendations are that a verification scheme's purposes explicitly include improving user safety; that verified status is visible by default; that an account only passes the filter if its user has been verified to a reliable standard; and that users are told both verification and the filter exist, and can easily find them.
We are critical of what has been proposed, and genuinely concerned that these proposals will fail to deliver the safety benefits which UK social media users have been repeatedly promised. But we've had some positive meetings with Ofcom officials, which we have appreciated, and we expect to have more. We'll do all we can to help Ofcom get these proposals to a good place.
Comments