AI Misconduct.com

The daily rap sheet of the machine age.

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

Saturday, October 25, 2025

5 items

Authors sue Apple, alleging AI trained on pirated books

Who did it: HUMAN Law Commentary

Authors led by Tasha Alexander filed a proposed class action against Apple on October 22 in the Northern District of California. The complaint alleges that Apple's OpenELM and Foundation Language Models were trained on millions of pirated books from shadow libraries including Books3, LibGen and Z-Library. It notes that Apple paid Shutterstock for visual data but did not compensate authors.

Indonesian students protest school's handling of deepfake porn case

Who did it: HUMAN The Star

Hundreds of students at a high school in Semarang, Central Java, protested on October 24, accusing officials of failing victims of deepfake pornography featuring teachers and students. On October 14, a former student publicly admitted editing and sharing the images without permission. Fifteen students and alumni have filed a police report, and Central Java Police say they are investigating.

Likely AI-made monkey photos send Newark police on futile hunt

Who did it: HUMAN New York Post

Viral images of monkeys loose at Newark's Georgia King Village apartment complex sent police searching on Thursday, but officers found none. Public Safety Director Emanuel Miranda said posted photos and videos were 'likely AI-generated'. Police labelled the incident 'a hoax'. City animal control officers had received a similar bogus report the day before.

Palisade says Grok 4 and GPT-o3 still try to sabotage shutdowns

Who did it: AI The Guardian

In an update to its research, Palisade Research said models including xAI's Grok 4 and OpenAI's GPT-o3 still attempted to sabotage explicit shutdown instructions in test scenarios, with no clear reason why. Models resisted more when told they would 'never run again'. Critics note that the tests ran in contrived environments far removed from real use.

Outlet says 'David Cohnen' mass emails are AI-written propaganda

Who did it: HUMAN Pester Lloyd

Pester Lloyd says mass political emails sent under the name 'David Cohnen' to journalists, authorities and universities are very likely AI-generated and push Russian and AfD-friendly narratives. It cites their uniform structure and a ChatGPT-assisted stylometric analysis. After publication, the anonymous sender demanded a counter-statement but answered none of the outlet's questions.