AI Misconduct.com

The daily rap sheet of the machine age.

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

Thursday, April 9, 2026

6 items

Florida opens OpenAI probe over claims ChatGPT helped FSU gunman

Who did it: HYBRID WCIV

Florida Attorney General James Uthmeier opened an investigation into OpenAI and ChatGPT following claims that the suspect in the FSU shooting, which killed two, used the chatbot to plan the attack. Uthmeier also cited national security concerns and alleged links to child sex abuse material, grooming and self-harm content, and said his office is issuing subpoenas.

Regulator says Dutch scan cars issue 500,000 unjustified fines yearly

Who did it: HYBRID NL Times

The Dutch Data Protection Authority calculates that municipal scan cars, which use AI algorithms to check parking payments, issue around 500,000 unjustified fines a year out of 3 to 5 million. The regulator says the cars can't tell parking from loading, can't always see disabled permits, and that not every municipality using them has assessed and mitigated the privacy risks.

Jakarta governor says AI-faked JAKI complaint reply was repeat offense

Who did it: HUMAN nasional

Jakarta Governor Pramono Anung said on April 9 that a JAKI app complaint in Kalisari closed with an allegedly AI-made photo was not the first such case, and that the same person was responsible. Pramono said the Kalisari village head and section staff had been removed; three PPSU officers face sanctions under their contracts, and the governor warned future offenders would be fired outright.

Louisiana bankruptcy lawyer fined $2,750 for AI-invented statute and citation

Who did it: HYBRID AI Hallucination Cases database

Bankruptcy Judge John W. Kolwe fined lawyer Kathleen M. Wilson $2,750 on April 9 after she admitted using generative AI for filings citing a nonexistent Louisiana statute and a case citation matching no opinion. She is barred from filing or joining bankruptcy cases in the Western District of Louisiana until she pays and completes six hours of in-person AI CLE; her filings were stricken.

Upstart Holdings hit with securities suit over overcautious AI model

Who did it: HUMAN The D&O Diary

A shareholder has filed a securities suit against Upstart Holdings alleging the AI lending platform failed to disclose that its Model 22 approval model overreacted to negative macroeconomic signals, making its guidance unreliable. Upstart had blamed Model 22 for missing 3Q25 revenue guidance on November 4, 2025, saying it "overreacted"; the shares fell about 10%.

Pro se litigant fined $5,000 over AI-fabricated case law

Who did it: HYBRID AI Hallucination Cases database

Chief Judge Virginia M. Kendall struck pro se plaintiff Ifeoma Delliane Chinedu Obi's motion to alter the judgment dismissing her suit against Cook County and fined her $5,000 under Rule 11, finding the AI-generated filings cited a nonexistent case, misquoted real ones and invented language from the court's own order; an earlier brief had at least 17 such fabrications.