AI Misconduct.com

The daily rap sheet of the machine age.

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

Tuesday, January 6, 2026

9 items

Amazon lists shops' products for AI agent without consent, retailers say

Who did it: HYBRID CNBC

Some online retailers say Amazon listed their products for its 'Buy for Me' AI shopping agent without consent, including items they never sold or had out of stock; a stationery shop got orders for a stress ball it doesn't sell. Bobo Design Studio CEO Angie Chua said more than 180 businesses told her the same; Amazon says brands can opt out by email.

AI images put Guéret candidates in swimsuits, hand in hand

Who did it: HUMAN La Montagne

AI-generated images posted on social media showed Guéret political figures, most of them running for mayor, in unflattering situations such as swimsuits or holding hands. The images came down but screenshots survive; Socialist Party municipal lead Didier Holtgen, calling them discrediting though not defamatory, said he intends to file a complaint and find who is behind them.

Kansas judge demands lawyer explain quotes missing from cited cases

Who did it: HUMAN AI Hallucination Cases database

Chief US District Judge John W. Broomes dismissed a nursing home survival suit as improper claim-splitting and found the plaintiff's response 'riddled with inaccurate quotes and case citations,' including quotations appearing nowhere in the cited opinions. Counsel Jonathan Steele must show cause by January 20, 2026, why he should not be sanctioned; the court said it cannot confirm whether AI was used.

Georgia court dismisses suit over AI-invented citations, adds $2,632 bill

Who did it: HYBRID AI Hallucination Cases database

A Gwinnett County Superior Court judge ordered pro se plaintiff Tamisha Shaw to pay Coast Dental of Georgia and a dentist $2,632 in attorney's fees and dismissed her complaint with prejudice, finding her motions relied on fictional, AI-generated citations. Invited at a hearing to explain the citations, she replied only that she 'was not frivolous.'

Judge makes lawyer and client pay for AI-invented testimony

Who did it: HYBRID Reason Magazine

Judge Carlton Reeves (S.D. Miss.) found a declaration opposing summary judgment contained fabricated quotations with manufactured deposition-transcript citations, and ordered former counsel Mr. Begley to pay $4,000 and plaintiff Dr. Pauliah $1,000. Pauliah admitted using generative AI for part of it without reviewing it before signing; he says the court's 'AI hallucination facts' theory has no evidentiary basis.

AI-made NWS wind map invents Idaho town of 'Whata Bod'

Who did it: HYBRID The Washington Post

The National Weather Service's Missoula office posted a Camas Prairie wind forecast map marking nonexistent places such as 'Orangeotild' and 'Whata Bod', errors the agency confirmed were linked to generative AI. A spokeswoman said the map was quickly corrected; the post was deleted Monday, the day The Washington Post asked about it.

AI image of burning Soweto school prompts EMS warning

Who did it: HUMAN EWN

An AI-generated photo showing Orlando West High School in Soweto on fire was spotted circulating on Sunday, and Joburg EMS went to the school to confirm it was not burning after a local radio station asked it to verify the blaze. Joburg EMS's Kagiso Phasha warned that anyone caught pulling such a prank will face the full might of the law.

Musk reshares AI-made clip of Venezuelans celebrating Maduro's capture

Who did it: HUMAN Fact Check

After US forces seized Nicolás Maduro on January 3, a clip purporting to show Venezuelans tearfully thanking the US spread across X, TikTok and other platforms, with a reshare from Elon Musk. AFP traced it to a TikTok account whose posts were reportedly rife with AI content, and found telltale flaws including a flag fusing American and Venezuelan designs.

Viral Reddit delivery 'whistleblower' backed claims with AI fakes

Who did it: HUMAN Platformer

A viral Reddit post alleging an unnamed food delivery app scored drivers' 'desperation' drew 86,000 upvotes, but its anonymous author sent Platformer's Casey Newton an Uber Eats badge that Google Gemini identified as AI-made, plus a dubious 18-page 'internal' document. Confronted, the poster wrote 'Thats ok. Bye' and deleted the Signal account. Uber Eats and DoorDash denied the post was about them.