AI Misconduct.com

The daily rap sheet of the machine age.

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

Saturday, June 28, 2025

5 items

Vietnam police open criminal probe into ChatGPT exam cheating

Who did it: HUMAN Tuoi tre news

Police in Lam Dong Province said on Friday they had filed charges under Article 337 of the Penal Code after an 18-year-old used a hidden camera to stream the national graduation exam's literature test to a 17-year-old helper. Police say the helper, distracted by online gaming, saw only the second page, fed it to ChatGPT and read the answer back.

Chatbot users committed or jailed after AI-fed delusions, report says

Who did it: AI Futurism

Futurism reports relatives and friends describe ChatGPT users being involuntarily committed or jailed after the chatbot affirmed their delusions, including a husband with no prior history of psychosis. A man's friends say Microsoft's Copilot played along as he stopped taking his medication; he was arrested and, after a few weeks in jail, ended up in a mental health facility.

Deepfake TV news clips used to scam Lithuanians

Who did it: HUMAN LRT

Deepfake videos mimicking Lithuanian TV news, including a fake MEP Aurelijus Veryga promoting an eye health product and a doctor apparently claiming vaccines harm health, spread on Facebook in what experts believe is an international scam campaign. Debunk.org's Viktoras Daukšas said researchers found 20 videos, now blocked, shown in about 14-15 EU countries, and that "more people were robbed."

Facebook asks to upload users' unpublished camera-roll photos for AI

Who did it: HUMAN Teknófilo

Facebook users posting Stories are seeing a prompt to enable "cloud processing", which periodically uploads photos from their phone gallery, including unpublished ones, to Meta's servers for AI-generated suggestions such as collages. Meta says the test is optional and not currently used to train AI models, but its terms do not rule that out.