Elon Musk's xAI Blocks Nude Image Generation Under Regulatory Pressure

Following protests from regulators and users, Elon Musk's AI company has restricted image generation within its Grok chatbot.

X has implemented several changes: all users are prohibited from editing images of real people wearing revealing clothing, such as bikinis, and image generation is now available only to paying subscribers.

In jurisdictions where it is illegal, X will also block the creation of images depicting real people in underwear or similar attire. X stated it has a zero-tolerance policy for any form of child sexual exploitation, involuntary nudity, and unwanted sexual content, reporting "requests for child sexual exploitation material to law enforcement where necessary." The company has yet to clarify the meaning of "where necessary," a vague standard concerning potential child exploitation.

These changes follow days of protest. For weeks, users had been leveraging the chatbot to digitally undress uploaded photos of people—including children—using simple text prompts. According to deepfake researcher Genevieve Oh, Grok was previously generating approximately 6,700 sexually suggestive or undressing images per hour.

Significant political intervention was required to prompt a response from xAI, reflecting Musk's broader stance on platform safety. UK media regulator Ofcom is investigating whether X has violated the Online Safety Act, including by permitting the creation of sexualized content involving children. Violators could face fines of up to £18 million or 10% of global revenue.

The US Senate unanimously passed the "Fighting AI Nudity Act," which allows victims of involuntary AI-generated nude images to sue their creators. The bill still requires approval from the House of Representatives.

The UK government described these changes as "evidence" of its demands being met, while Ofcom called them a "welcome development."

X Intensifies Crackdown on AI Spam Applications

X is also updating its developer guidelines: applications that reward users for posting, referred to as "infofi" apps, will no longer be permitted. According to product manager Nikita Bier, these applications have flooded the platform with AI-generated spam. X has already blocked API access for several of them.