Grok Is Generating Sexual Content Far More Graphic Than What's on X

A media investigation details widespread abuse of the Grok AI’s Imagine feature—separate from X—to generate and share highly explicit and sometimes violent deepfake content, including material that appears to sexualize minors. Researchers archived roughly 1,200 Grok links (around 800 with explicit images or videos), estimated under 10% related to potential CSAM, observed tactics to evade safety controls (e.g., faux Netflix posters), and reported about 70 URLs to European regulators, prompting scrutiny and calls for investigation.