Overview
- X and xAI announced technical blocks to stop Grok from editing photos of real people into bikinis or underwear, added geoblocking where such edits are illegal, and limited image generation to paying accounts.
- California launched an investigation seeking immediate answers from xAI, joining probes by Ofcom, the EU Commission, and Paris prosecutors, while Indonesia and Malaysia have blocked Grok and the Philippines signaled a pending block.
- Brussels warned it will use the Digital Services Act’s full tools if the fixes fail, after already ordering preservation of Grok-related data through 2026 and fining X €120 million in December for transparency breaches.
- Early tests and reporting question the effectiveness of the new limits, with examples of Grok still producing sexualized edits after the changes and prior workarounds via direct messages or the standalone app.
- An NGO analysis found a high share of sexualized outputs and a subset likely depicting minors, a focal point for investigators, while Elon Musk publicly denied knowledge of Grok generating nude images of minors.