AI Misconduct.com

The daily rap sheet of the machine age.

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

Tuesday, December 9, 2025

8 items

Brussels opens antitrust probe into how Google feeds its AI

Who did it: HUMAN European Commission

The European Commission opened a formal antitrust investigation into whether Google abused a dominant position by using web publishers' content for AI Overviews and AI Mode, and YouTube videos to train its AI models, without appropriate compensation or a way to refuse. The Commission says rival developers are barred from training on YouTube content, and that opening a probe does not prejudge the outcome.

CHA law firm and lawyer sanctioned over ChatGPT-invented case

Who did it: HYBRID Chicago Sun-Times

Cook County Circuit Judge Thomas Cushing sanctioned Goldberg Segalla $49,500 and lawyer Larry Mason $10,000 after a post-trial motion defending the Chicago Housing Authority in a lead-paint case cited a fake court case. Danielle Malaty, the attorney who used ChatGPT and has said she didn't think it could invent citations, was not penalized; the judge also threw out the motion.

Litigant's AI-drafted reply cites fake cases in India's Supreme Court

Who did it: HYBRID The Times of India

In what the report calls the first such case detected in India's Supreme Court, a rejoinder drafted with AI tools for Gstaad Hotels Bengaluru promoter Deepak Raheja, in a dispute with Omkara Asset Reconstruction, cited hundreds of fake cases. His senior advocate admitted the mistake, saying he had never been more embarrassed; the court said it could not brush it aside.

Mexican Senate leader rebukes Google over Sheinbaum deepfake investment ads

Who did it: HUMAN La Jornada

Adán López Hernández, head of the Mexican Senate's Political Coordination Board, sent Google México a letter rebuking the company for allowing ads that use AI to impersonate President Claudia Sheinbaum and solicit investments, for example in Pemex. He argues the ads' 'sponsored' label means someone paid for them with Google's consent and to its profit, and says they have intensified on YouTube.

Deepfake Sitharaman video costs Bengaluru woman over Rs 33 lakh

Who did it: HUMAN The Times of India

According to a complaint to Bengaluru's East Cyber Crime police, a woman joined a trading platform after seeing a deepfake Facebook video of finance minister Nirmala Sitharaman promoting it. Told months after investing Rs 21,000 that her account showed an $88,000 profit, withdrawable only after taxes and fees, she sent Rs 33,25,413 in nine transfers. Police have registered a cheating case.

Waymo to recall software after robotaxis passed stopped school buses

Who did it: AI USA Today

Waymo says it will file a voluntary software recall with NHTSA after a robotaxi passed a stopped school bus and students in Atlanta on September 22. Austin's school district says Waymo vehicles passed buses or students at least 20 times since August 2025, with at least five violations after Waymo said updates had fixed the issue; Waymo denied its November 20 request to halt service.

AI hoax image of bridge damage after quake disrupts 32 UK trains

Who did it: HUMAN Futurism

After a 3.3 magnitude earthquake in Lancashire, an AI-generated image appearing to show the 179-year-old Carlisle Rail Bridge partly collapsed spread on social media, the BBC reported. Rail officials halted trains over the undamaged bridge while it was inspected; 32 passenger and freight services were delayed or halted, some as far north as Scotland.

Radnor High School and police investigate AI video of students

Who did it: HUMAN 6abc Philadelphia

Radnor High School and Radnor Township Police are investigating an AI-generated video that was reported to depict several students 'in an inappropriate manner', according to an email to families from principal Joseph MacNamara. The school says the students' families have been contacted and offered support; police would say only that the incident is under active investigation.