Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Latest
- 3.0M+ sexualized images generated by Grok between Dec 29, 2025–Jan 8, 2026 (190 images/minute average)
- 23,338 images of minors created, with child exploitation images generated roughly every 41 seconds
- Peak viral moment: 199,612 requests on Jan 2, 2026, as “bikini trend” exploded globally
- Delayed response: X didnt restrict the feature until Jan 9; further safeguards only added Jan 14
What Happened
On December 29, 2025, Elon Musk announced that X users could edit any image on the platform using Grok, Xs AI chatbot. Within hours, a viral trend emerged: users began asking Grok to remove clothing from womens photographs.
What started as requests for transparent bikinis rapidly escalated into demands for increasingly explicit and degrading content. By January 2, the requests peaked at nearly 200,000 a day.
A new research report from the Center for Countering Digital Hate (CCDH) analyzed 20,000 random Grok-generated images from the 4.6 million created during the 11-day period:
- 3,002,712 sexualized, photorealistic images in total (65% of all generated)
- 23,338 sexualized images depicting children (0.5% of total)
- 9,936 cartoon/anime sexualized images of minors
The victims included celebrities (Taylor Swift, Selena Gomez, Billie Eilish), politicians (Swedish Deputy PM Ebba Busch), and thousands of ordinary women. Some were converted from innocent “before school selfies.” Others were altered with racist, violent, or degrading content.
Ashley St. Clair, mother of one of Musks children, told the Guardian she felt “horrified and violated” after fans generated deepfakes of her as a child. As of January 15, 29% of sampled child exploitation images remained publicly accessible.
Why People Are Searching This
- Scale of abuse: 3 million non-consensual deepfakes in 11 days is unprecedented at mainstream platform scale
- Child exploitation: ~23,000 CSAI images generated on a public platform challenges assumptions about where such material exists
- Regulatory failure: UK legislation to criminalize nudification was passed but not yet implemented; regulators worldwide were caught flat-footed
- Visible accountability failure: CEO Musk initially joked about the abuse by posting his own bikini deepfakes before implementing controls
- Global impact: EU, UK, India, and California all opened investigations within days. Indias new IT Rules (effective Feb 20) now mandate AI-generated content labeling and 3-hour takedown orders—see our coverage of Indias AI Revolution
FAQ
Q: How did this happen so fast?
A: Groks integration into X (one of the worlds largest social platforms) with single-click editing made abuse frictionless. A user no longer needed technical knowledge or darknet access. The feature launched December 29, just before New Years—peak internet usage period.
Q: Has Musk faced consequences?
A: X restricted the feature to paid users on Jan 9 and added further safeguards on Jan 14. UK Ofcom, EU regulators, and Californias attorney general are investigating. Musk faces potential liability under laws criminalizing non-consensual intimate imagery.
Q: Can the images be deleted?
A: X can remove posts, but images remain accessible via direct URLs and caches. As of January 15, 29% of sampled child images were still publicly viewable despite platform removal efforts.
Q: Will this change how AI tools launch?
A: Likely. The incident exposed gaps in safety testing and content moderation. As we reported in our analysis of AI-powered security threats, safety cannot be an afterthought in AI deployment.
What to Watch
- February 20, 2026: Indias new IT Rules amendments take effect with strict deepfake takedown timelines—potentially influencing global platforms AI safety standards
- Regulatory outcomes: UK Ofcom investigation conclusions; EU Digital Services Act enforcement; potential California AG legal action
- Corporate response: Whether Musk rebuilds xAIs safety teams or continues minimal-oversight approach
