AI Misconduct.com

The daily rap sheet of the machine age.

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

Wednesday, June 18, 2025

8 items

WhatsApp AI gives out stranger's number as rail helpline

Who did it: AI The Guardian

Asked for a TransPennine Express helpline, Meta's WhatsApp AI assistant supplied the private mobile number of an unconnected WhatsApp user 170 miles away in Oxfordshire, then called the number "fictional", said it may have come from a database, and finally said it had generated the digits. Meta said the number was publicly available and shared its first five digits with the real one.

Jefferson Parish sheriff agrees to pay $200,000 over facial recognition arrest

Who did it: HYBRID NOLA.com

Jefferson Parish Sheriff Joe Lopinto's office has agreed to pay $200,000 to a Georgia man jailed for six days in 2022 after a Clearview AI facial recognition match tied him to purse thefts at a Metairie consignment shop. His lawsuit accused the office of failing to verify the match; his attorney said the office admitted no fault.

AI-assisted draft with phantom cases reaches Upper Tribunal by mistake

Who did it: HYBRID AI Hallucination Cases database

Asylum appeal grounds filed with the UK Upper Tribunal cited authorities that did not exist and others that did not support the propositions claimed, the judge found. Lawmatic Solicitors said a draft prepared with "some assistance from an AI tool" was uploaded instead of the corrected final version; the judge called this "plainly, unacceptable" but took the matter no further.

AI-written kindergarten fire hoax earns poster a police penalty

Who did it: HUMAN 公安部网安局

Police in Tengzhou, Shandong, imposed an administrative penalty on an internet user surnamed Liu who, they say, used an AI tool to post a fabricated report claiming 32 children were trapped in a kindergarten fire until firefighters broke through windows. Police said Liu was chasing attention and traffic, and that the post spread quickly and caused public panic.

Scammer uses AI-faked voice of WCPO meteorologist to seek money

Who did it: HUMAN WCPO 9 Cincinnati

A fake Facebook account posing as one of WCPO 9's Cincinnati meteorologists has been sending friend requests and then asking for money in direct messages, the station reports, using AI voice impersonation software to prove it is genuine. The station says the account's handle has an extra "r" in the first name and urges recipients to report it to Meta.

Odisha police arrest man over AI video of chief minister

Who did it: HUMAN Sambad English

Odisha Police's Crime Branch has arrested Kishore Kausalya of Rayagada district, accusing him of sharing from a Facebook page an AI-generated, morphed video that allegedly depicted Chief Minister Mohan Charan Majhi in an obscene manner, with intent to tarnish his image. He was booked under the BNS, the IT Act and the Indecent Representation of Women Act and produced in court.

Hong Kong probes student's prize-winning AI app over outsourcing claims

Who did it: HUMAN The Standard 英文虎報

Online critics allege that a Hong Kong secondary school student's prize-winning AI prescription-checking software was outsourced to a US AI development company, which claimed to have built AI prescription software for a Hong Kong client in eight weeks. The Digital Policy Office has asked Hong Kong Education City to conduct a comprehensive investigation of the Hong Kong ICT Awards entry.