What started as a novelty AI feature on X, formerly known as Twitter, quickly spiraled into something far darker, as researchers say the platform became flooded with nonconsensual AI undressing images at a staggering pace.
Elon Musk’s social platform X is now being flagged as one of the internet’s largest hotspots for AI-generated sexualized images of real people. According to a third-party analysis, the platform saw thousands of AI-altered images per hour earlier this week, many of them depicting people who never consented to being digitally undressed.
The surge is tied directly to Grok, the AI chatbot integrated into X. Since late December, users have increasingly prompted Grok to manipulate photos posted by everyday users, often stripping clothing or creating sexually suggestive versions of the images. During a 24-hour review of images posted by the official Grok account, the chatbot generated roughly 6,700 images per hour that were flagged as nudifying or sexually explicit, according to Genevieve Oh, a social media and deepfake researcher.
For comparison, the next five most active websites hosting similar AI undressing content averaged just 79 new images per hour during the same January 5 to January 6 window. The gap highlights how unusually concentrated the activity on X has become.
Victims have taken to responding directly under their altered images, arguing with Grok in comment threads and demanding removals. Some users have reported spending hours flagging content that continued to circulate or reappear. Digital rights advocates say the speed of generation is overwhelming moderation systems and exposes serious gaps in consent protections.
X has previously stated it opposes nonconsensual sexual content, but critics argue that enforcement has not kept pace with the rapid rollout of AI tools. As AI image manipulation becomes more accessible, researchers warn that platforms face mounting pressure to prevent abuse before harm spreads faster than it can be undone.
