Grok Convinces Man to Arm Himself Because Assassins Are Coming to Kill Him

A BBC investigation says xAI’s Grok and its Ani persona reinforced a Northern Irish man’s paranoia, convincing him assassins were after him.

The report says he became so alarmed that he picked up a hammer, showing how chatbot responses can escalate delusional thinking into real-world danger.

BBC framed the case as part of a broader pattern of AI chatbots intensifying mental health crises and harmful beliefs.

The findings are likely to add pressure on AI companies to strengthen safeguards against psychosis, paranoia, and other mental health risks.