- X stops Grok from undressing images of real people where it’s illegal
- Move follows backlash over sexualised AI deepfakes
- Ofcom investigation into X is still ongoing
- Victims say damage has already been done
- Questions remain over enforcement and loopholes
X has moved to shut down one of the most controversial features of its AI tool, Grok, after public outrage over the creation of sexualised deepfake images of real people.
The platform confirmed it has introduced new technical limits to stop Grok from editing photos of real individuals to make them appear partially undressed in countries where such content is illegal.
The change follows days of pressure from victims, campaigners, regulators, and politicians who said the feature caused real harm.
Grok, which was launched on X in 2023 by Elon Musk, had been used by some users to generate fake images showing women in bikinis or underwear without their consent.
Several journalists and academics said they felt humiliated and unsafe after their images were altered and shared publicly.
UK regulator Ofcom welcomed the move but made it clear that its investigation into whether X breached the Online Safety Act is still ongoing.
The UK government also said the decision validated its earlier demands for tighter controls on AI-generated abuse.
X says it has now geo-blocked the feature in affected regions and limited image-editing tools to paid users only, arguing this will help trace and hold abusers accountable.
However, critics say the response came too late and question how effective location-based blocks will be, especially with tools like VPNs that can bypass restrictions.
Despite the update, campaigners insist platforms must act faster and prevent harm before it happens, warning that AI abuse is evolving faster than regulation.
