AI Misconduct.com

The daily rap sheet of the machine age.

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

Wednesday, March 25, 2026

7 items

Canadian immigration refusal cites robot-panel job duties applicant never listed

Who did it: HYBRID Toronto Star

Canada's Immigration Department rejected a McMaster University post-doctoral fellow's permanent residence application, citing duties such as wiring control circuits and building robot panels that she never claimed. The letter disclosed that generative AI supported processing; her lawyer says it 'hallucinated' her job description. The department says a human officer made the decision. The file has been reopened.

Judge says Pentagon blacklisting looks like punishing Anthropic

Who did it: HUMAN Reuters

At a hearing on Anthropic's bid to block its designation as a national security supply-chain risk, US District Judge Rita Lin said the move "looks like an attempt to cripple Anthropic", which had refused to let the military use Claude for US surveillance or autonomous weapons. The government argues it could not rely on the company; Lin will rule in writing.

Teens get probation over AI fake nudes of classmates

Who did it: HUMAN Associated Press

Two teenage boys, 14 at the time, who admitted making about 350 AI-generated fake nude images of at least 59 girls under 18, many of them classmates at a Pennsylvania private school, were given probation, 60 hours of community service and unspecified restitution. The judge said that as adults they would probably be headed for state prison.

Teen jailed for life after tricking DeepSeek over murder plan

Who did it: HYBRID Metro

A teenager who admitted murdering his mother in north Wales was jailed for life with a minimum term of 22 and a half years. Mold Crown Court heard he had asked DeepSeek for tips for a 'non-experienced' killer and, when it refused to engage, tricked it into responding by claiming he was writing a book about serial killers.

German court awards student €200 over AI exam face checks

Who did it: HUMAN dpa

Germany's Higher Regional Court in Jena ruled that a university's AI-supported proctoring software, which automatically compared webcam shots taken during online exams with reference photos, processed biometric data in breach of the GDPR's Article 9 ban. Overturning a dismissal in Erfurt, it awarded a student €200 for a diffuse feeling of being watched and fear of being wrongly suspected of cheating.

Ex-nutritionist claims Real Madrid medics prescribed supplements via ChatGPT

Who did it: HUMAN Canary

A former Real Madrid nutritionist claimed on Instagram that some medical staff decisions were based on free AI programs, wondering whether scanning the wrong knee, as journalist Miguel Ángel Díaz alleges happened with Kylian Mbappé, is worse than "prescribing supplements based on free Chat-GPT". Mbappé says his knee is fine and that "some false things have been said".

Thai actor Pond reports AI porn fakes sold in secret groups

Who did it: HUMAN amarintv.com

Pond, a 25-year-old actor from the film "Phi Nak 5", filed a police complaint at Thailand's Central Investigation Bureau on 25 March, saying scammers used AI to edit his face and private chats into pornographic clips billed as his leaked videos and sold through secret groups for about 800–1,000 baht. When he posted a warning, he says, they began offering access free.