AI Misconduct.com

The daily rap sheet of the machine age.

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

Friday, April 25, 2025

8 items

Judge orders Lindell's lawyers to explain AI brief's phantom cases

Who did it: HYBRID kdvr.com

Federal judge Nina Wang found about 30 defective citations, including cases that do not exist, in a brief Mike Lindell's lawyers filed in February in a former Dominion executive's defamation case, and ordered Lindell, attorney Christopher Kachouroff and Jennifer DeMaster to show why they should not be sanctioned. The team admits using AI but says an earlier draft was filed by "human error".

Udayana University student accused of making AI deepfake porn

Who did it: HUMAN Kompas.com

A Universitas Udayana student in Bali is suspected of using AI deepfake technology to create and spread fake pornographic content, a case that went viral on X. A university official said the student is barred from academic and administrative services pending an ethics hearing on sanctions; the number of victims is not yet known.

California State Bar admits AI helped write bar exam questions

Who did it: HYBRID Inside Higher Ed

The State Bar of California said some multiple-choice questions on its February bar exam were developed by ACS Ventures, its independent psychometrician, using AI, after test takers complained of glitches. The bar plans to ask the California Supreme Court to adjust scores but defends the questions; a law professor called it a "staggering admission".

AI audio tool helped Israel target Hamas commander, officials say

Who did it: HYBRID The New York Times

According to Israeli and American officials, Unit 8200 engineers added AI to an audio tool that gave an approximate location for Hamas commander Ibrahim Biari, and Israel struck the area on Oct. 31, 2023, killing him and, per the monitor Airwars, more than 125 civilians. Officers had been warned the area was densely populated. Israel's military says it is investigating the strike.

xAI accused of running unpermitted gas turbines for Memphis supercomputer

Who did it: HUMAN The Guardian

The Southern Environmental Law Center says Elon Musk's xAI moved at least 35 methane gas turbines without air permits into its Memphis site powering the Colossus supercomputer, and thermal images show 33 likely running; the mayor said only 15 were on. Anonymous fliers telling residents the turbines are "specially designed to protect the air" arrived before a health department hearing.

Paraná prosecutors sue state over facial recognition in schools

Who did it: HUMAN Folha de S.Paulo

Paraná's Public Prosecutor's Office has sued the state government and contractor Valid Soluções, seeking R$15 million in collective moral damages and a halt to collecting biometric data from nearly 1 million children and adolescents via facial recognition attendance-taking used since 2023. The prosecutor says consent forms did not specify facial biometrics or offer refusal; the state says the tool follows privacy laws.

B.C. tribunal makes litigant pay $500 over AI-invented case law

Who did it: HYBRID AI Hallucination Cases database

B.C.'s Civil Resolution Tribunal dismissed a self-represented claimant's service-dog claim against gas station owner Hung Long Enterprises and ordered her to pay it $500 for its time, partly over AI-assisted submissions citing a non-existent CRT case and three non-existent Supreme Court of Canada cases. She admitted using AI and apologised but stood by her arguments.

Deepfake FT writer pitches fake investment group in Meta ads

Who did it: HUMAN Financial Times

Deepfake videos and doctored images of an FT writer, promoting a WhatsApp investment group the writer supposedly leads, ran in over 1,700 Facebook and Instagram ads that reached over 970,000 EU users, according to Meta's Ad Library data cited by the FT. Ten fake accounts bought them; Meta says it removed the ads, accounts and pages reported to it.