Elon Musk-owned X is reviewing reports of suspected illegal hate and terrorist content faster in the UK, but regulator Ofcom says it remains concerned about whether the platform is actually removing illegal material effectively.
Ofcom said on October 14 that X’s first quarterly performance report showed improvements in the average time taken to review reported material compared with evidence examined earlier in 2026. The company’s performance also fell within previously agreed response-time commitments.
Reuters independently reported that the regulator nevertheless has ongoing concerns about the effectiveness of X’s efforts to take down illegal hate and terror content.
Faster review does not necessarily mean effective moderation
In May, X committed to reviewing suspected illegal terrorist and hate content reported through its dedicated UK tool within an average of 24 hours over a three-month period. It also agreed to assess at least 85% of such reports within 48 hours.
The latest submission indicates that X is meeting those timing commitments.
But Ofcom is now examining what happens after the review.
The regulator said it will cross-check X’s performance figures through its own testing and has received mixed reports about the company’s sustained engagement with third-party experts.
That distinction matters for platform engineering because moderation systems can optimize operational metrics such as review time without necessarily improving the final outcome for users.
Grok restrictions have expanded in the UK
Ofcom’s update also includes new developments in its investigation of Grok, the AI service originally developed by xAI and now operated by SpaceXAI.
The regulator opened its formal X investigation in January after reports that Grok was being used to create and share sexualized images that could amount to intimate image abuse, pornography or child sexual abuse material.
Ofcom now says SpaceXAI has introduced additional restrictions for the standalone Grok service.
People using UK IP addresses have been restricted from viewing, generating or sharing pornographic content through Grok’s image-generation technology. Companion chatbots including Ani have also been restricted from engaging in sexually explicit conversations with UK users.
Reuters reported that Ofcom is continuing to determine whether X has fulfilled its legal duties to protect people in Britain from illegal content. X did not immediately respond to Reuters’ request for comment.
Online safety is becoming a systems-performance problem
The update exposes a difficult measurement problem for platforms.
Moderation teams can count how quickly a report enters a queue, how quickly somebody reviews it and whether a decision is recorded. Those metrics are relatively straightforward.
Measuring whether harmful material actually stops reaching people is harder.
For product and trust-and-safety teams, Ofcom’s scrutiny therefore signals that regulators may increasingly look beyond response times toward actual moderation outcomes.
That changes the engineering target.
A safety system cannot be judged only by how fast it processes reports if illegal material remains available afterward. As regulation becomes more operational, platforms may have to prove not simply that their moderation pipelines are running quickly—but that they are producing results that meaningfully reduce harm.
