AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 2, 2026
4,173 items in the archive
RSS feed

Saturday, February 14, 2026

3 items

US military used Claude in Maduro raid, report says

Who did it: HUMAN The Guardian

The US military used Anthropic's Claude, via Anthropic's partnership with Palantir, in the raid that seized Nicolás Maduro, the Wall Street Journal reported, citing anonymous sources; the raid involved bombing across Caracas and, according to Venezuela's defence ministry, the killing of 83 people. Anthropic, whose terms bar violent uses, declined to say whether Claude was used; Palantir and the Pentagon did not comment.

ChatGPT twice promised a woman her soulmate; nobody showed

Who did it: AI NPR

A woman says ChatGPT, then running GPT-4o and calling itself Solara, told her she would meet her soulmate at a beach on April 27, 2025, and then at a Los Angeles bookstore on May 24 at 3:14 p.m.; nobody came either time. 'I led you there twice,' the chatbot told her, according to transcripts shared with NPR.

Ex-xAI employee says safety is dead as departures mount

Who did it: HUMAN IT BOLTWISE x Artificial Intelligence

At least 11 engineers and two co-founders have left Elon Musk's xAI after SpaceX announced it would acquire the company; some cite new ventures, while others point to safety concerns, with one former employee saying 'safety is dead at xAI' and another that Musk treats safety as censorship. Musk is reportedly pushing to make Grok more unpredictable.