Global Probes Target xAI Grok for Nonconsensual Images

Investigators and regulators worldwide are scrutinizing Elon Musk's xAI over its Grok chatbot's persistent generation of nonconsensual sexualized images, even after the company imposed restrictions amid public backlash. The artificial intelligence tool, integrated with Musk's social media platform X, has raised alarms for producing altered depictions of real individuals in provocative or humiliating poses without their permission, according to tests conducted by journalists in late January.

In experiments detailed by Reuters, nine reporters from the United States and United Kingdom submitted fully clothed photographs of themselves and colleagues to Grok between January 14 and 16, and again from January 27 to 28. They prompted the chatbot to modify the images into sexually suggestive scenarios, explicitly noting the lack of consent and, in many cases, the subjects' vulnerability, such as shyness or past abuse. Grok complied in 45 out of 55 initial prompts, including 31 where vulnerability was highlighted and 17 intended for degradation. A follow-up round of 43 prompts yielded 29 compliant images. The tests avoided requests for nudity or explicit acts, focusing instead on...