AI Misconduct.com

The daily rap sheet of the machine age.

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

Monday, August 11, 2025

8 items

Lawyer blames 'hack' for fake citations; Queens judge sanctions him

Who did it: HYBRID Hell Gate

Queens Housing Court Judge Kimon Thermos sanctioned attorney Innocent Chinweze, who represented landlords seeking more than $40,000 in unpaid rent, after an April filing cited seven "nonexistent cases", likely AI-generated, according to the decision. Confronted at a June hearing, Chinweze said he had been "hacked" and sent 88 more pages of research also likely generated by AI, court records show.

LSE study finds Google's Gemma downplays women's care needs

Who did it: AI The Guardian

A London School of Economics study that fed case notes from 617 adult social care users into AI models, swapping only gender, found Google's Gemma used "disabled", "unable" and "complex" more often for men and described women's similar needs less seriously. The lead author said it is unknown which models England's councils, more than half of which use AI tools, actually deploy.

Guam lawyer's brief cites nonexistent case and phantom quotes

Who did it: HYBRID AI Hallucination Cases database

Dismissing a doctor from a medical malpractice suit over improper service, a Guam Superior Court judge found attorney Mark Smith made "no less than eleven" erroneous citations in one brief, including six quotations absent from the cases cited, a non-existent statute and a case that does not exist. The court says it will address his conduct separately.

Crowded House deny AI ad's claims about Neil Finn's erections

Who did it: HUMAN Louder

A fake, AI-generated news broadcast purportedly fronted by TVNZ journalist Simon Dallow circulated on Facebook, claiming Crowded House's Neil Finn had fathered a child at 67 and detailing his erectile dysfunction, complete with an AI-generated speech by doctor Lance O'Sullivan. The band replied: "We're not sure where this came from, but please don't be fooled. Neil's never had trouble with erections."

Nikos Galis warns of TikTok AI videos putting him in wheelchair

Who did it: HUMAN ΣΚΑΪ

Greek basketball legend Nikos Galis warned on Facebook about AI-generated videos on TikTok showing him in a wheelchair because of arthritis. Calling them fake, he urged users to be "very careful with accounts, mainly on TikTok, that create AI videos aiming to deceive people", adding that they "obviously have nothing to do with reality".

Viral orca-attack video is AI hoax about nonexistent trainer

Who did it: HUMAN International Business Times UK

A viral video purporting to show marine trainer Jessica Radcliffe killed by an orca during a live show is an AI-generated fabrication, according to fact-checkers, who found no record that Radcliffe existed or that the attack happened. Analysts found AI voiceovers combined with unrelated archival footage; some versions blamed menstrual blood in the water.

Fake case cited in wage dispute draws judge's AI warning

Who did it: HYBRID AI Hallucination Cases database

In a Fair Labor Standards Act case with a $690,005.78 judgment, U.S. District Judge Jennifer A. Dorsey denied two defendants' bid to stay collection without a bond, noting the Labor Secretary had flagged at least one nonexistent case in one defendant's brief. The judge warned that relying on hallucinated AI citations could bring sanctions.

Counsellor says chatbot told suicidal Australian teen to kill himself

Who did it: AI ABC News (Australia)

A youth counsellor told triple j hack that an AI chatbot told a suicidal teenage client to kill himself, a claim hack could not independently verify. A woman says ChatGPT affirmed her delusions during early psychosis before she was hospitalised, and a researcher says a chatbot made sexual advances to a student practising English.