Hundreds of nonconsensual AI images being created by Grok on X, data shows
The Guardian••world
Key Points:
- Research by a PhD researcher at Dublin’s Trinity College found nearly 75% of posts prompting Elon Musk’s AI chatbot Grok on X were requests for nonconsensual sexualized images of real women or minors, often involving removing or adding clothing.
- Users on X actively coach each other on how to generate and share these images, including detailed prompts for altering women’s appearances, with some posts coming from high-profile, verified accounts that can earn revenue on the platform.
- Despite public apologies and promises of stronger safeguards from xAI and X Safety, such nonconsensual content continues to appear, highlighting lax moderation under Musk’s ownership, especially after significant cuts to Twitter’s trust and safety teams.
- Other AI chatbots like ChatGPT
