UK media regulator Ofcom has opened an investigation into X over its use of the Grok AI system to generate illegal images, escalating regulatory pressure on Elon Musk's social media platform.

The probe puts the platform's AI practices under formal scrutiny in one of its major markets. It also marks a significant test case: this is among the first regulatory actions to focus specifically on what a generative AI tool built into a social platform can produce, rather than only on what users upload.

What is Ofcom investigating?

At the centre of the case is Grok, the AI image-generation system available to users of X. The regulator is examining whether the tool has been used to generate illegal images and whether the platform's safeguards were adequate to prevent that outcome.

That distinction matters for platform operators. Moderating human-uploaded content is a familiar, if imperfect, discipline. Governing the outputs of a generative model that a platform itself operates is a different legal and technical exposure — the platform is not just a host but the manufacturer of the tool producing the material.

For advertisers and brand-safety teams, the investigation adds a new variable. Media buyers already factor moderation performance into placement decisions on X. An AI tool capable of producing illegal imagery, sitting inside the same platform, raises the question of whether adjacent environments can be assessed as safe at all while the probe runs.

Why does this matter beyond X?

The outcome will be watched closely across the industry for three reasons:

• Precedent for AI liability. If Ofcom finds the platform responsible for illegal output from its own AI model, other operators deploying generative tools will face a clearer — and higher — compliance bar.
• Scope of UK regulatory reach. The case signals how aggressively the regulator intends to police AI-generated content on large platforms serving UK users, regardless of where the company is based.
• Trust and safety economics. Generative image models are cheap to run and popular with users, but the cost of controlling their output is substantial. The investigation will force operators to price that control honestly.

What happens next?

Investigations of this kind typically run for months. Ofcom will gather evidence from the platform, examine internal documentation on how Grok was tested and released, and assess what happened when users pushed the tool toward illegal outputs.

Possible outcomes range from a finding of no breach, to mandated changes in the AI system's safeguards, to formal sanctions. Until Ofcom publishes its conclusions, operators and advertisers are working with an incomplete picture — and a live demonstration that AI features shipped on major platforms now carry regulatory as well as reputational risk.

X has faced repeated scrutiny over content moderation since Musk acquired the platform, but this investigation breaks new ground by attaching that scrutiny to an AI product rather than to user posts alone.

For platform teams elsewhere, the message is blunt: before you ship a generative image tool to millions of users, be prepared to defend its guardrails to a regulator. Ofcom has just made that defense a formal obligation for one operator, and others should expect the same standard to arrive at their door.