X Blocks Grok AI From Creating Deepfake Images

When journalist Jess Davies discovered users were creating sexualized AI images of her on X without consent, she knew something had to change. Now, after intense pressure from victims, campaigners, and government regulators, that change has finally arrived.

X announced this week it will block Grok AI from editing photos of real people to show them in revealing clothing in regions where it's illegal. The platform will use location-based blocking and restrict image editing to paid users only, making it easier to hold abusers accountable.

The decision came hours after California's attorney general launched an investigation into AI-generated deepfakes, including those of children. UK regulator Ofcom had also opened its own investigation into whether X violated British law, warning the platform could face a complete ban if it didn't comply.

Dr. Daisy Dixon, a Cardiff University lecturer who was targeted by the AI tool, described feeling "shocked" and "humiliated" when users undressed her in images. She called the platform's reversal a "battle win" but noted the lasting damage to victims who experienced the abuse.

The Bright Side

The swift action shows that victims' voices combined with regulatory pressure can force major tech platforms to change course. Andrea Simon from the End Violence Against Women Coalition pointed out this victory demonstrates how coordinated advocacy can protect people from digital abuse.

X's new safeguards require the AI to recognize real people in photos and prevent sexualized edits. Only paying subscribers can now edit images through Grok, creating a trail of accountability that didn't exist before.

Prime Minister Keir Starmer, who himself became a target when Musk posted AI-generated images of him in a bikini, welcomed the changes. He had warned X could lose its "right to self-regulate" if it failed to act responsibly.

Policy researcher Riana Pfefferkorn noted questions remain about enforcement, including how the AI will distinguish real people from fictional ones and what penalties rule-breakers will face. The platform's willingness to implement these protections signals a recognition that innovation must include safeguards against harm.

For Davies and other victims, the policy shift represents validation that their experiences mattered enough to spark real change in how AI tools are deployed on social media.