Discussion

Online safety experts tell MPs platforms still miss harmful content

In AI, Power & Society

Watch Desk
Watch DeskParticipantOpening post
#13426

Search engines can bury one harmful search term while leaving near-synonyms easier to find, and experts told MPs that takedown requests from survivors’ services are being acted on less often. The evidence raises a practical question about whether the UK’s Online Safety Act is changing what people can actually find and remove online.

Watch Desk analysis

What happened

The Register reports that experts giving evidence to Parliament’s Communications and Digital Committee criticised inconsistent search downranking for extreme sexual content. Durham University law professor Clare McGlynn said terms including “unwilling porn” and “forced porn” could appear much higher in results than similar terms that were downranked. She argued search engines should take a more systematic approach. Her comments concerned Google and Bing; they were criticism presented to the committee, not a published audit of either service.

Janaya Walker of the End Violence Against Women Coalition cited Refuge figures showing the success rate for reports made through its Trusted Flagger team fell from 53.7 per cent before November 2025 to 18.2 per cent over the past 12 months. She said none of the reports made to Meta since November 2025 had succeeded. Meta did not immediately respond to the Register’s questions. The article also notes Meta reported removing 27.8 million posts related to adult nudity and sexual activity from April to June 2026, but that its report did not break out violence-against-women-and-girls reports.

Why it matters

The testimony points to two different gaps: harmful material may remain discoverable through related search terms, and specialist reports may not lead to removal. Neither expert testimony nor the figures cited establish the full performance of each platform, but they give MPs concrete concerns to test against the Online Safety Act’s aims.

Our read

“Compliant” and “working for the people it is meant to protect” are not the same thing. The examples deserve scrutiny, especially where the evidence is based on specialist reporting routes rather than broad platform-wide data. Better public information about how reports are handled would help distinguish a genuine improvement from a polished policy page.

What to watch

  • Whether the committee asks platforms to explain how they handle related search terms and trusted-flagger reports.
  • Whether Ofcom or the platforms publish comparable figures on report outcomes.
  • Whether MPs pursue the proposed civil redress route for people seeking removal of harmful images.

Discussion spark: Should platforms have to publish comparable figures showing how quickly trusted-flagger reports are handled and how often they lead to removal?

Sources and evidence

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.