AI Misconduct.com

The daily rap sheet of the machine age.

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

Tuesday, March 24, 2026

8 items

Israeli brothers indicted for selling Iran AI-faked intelligence

Who did it: HUMAN The Jerusalem Post

Two brothers were indicted on charges including contact with a foreign agent, accused of taking more than NIS 100,000 from Iranian agents for information that, the indictment says, was largely fabricated with help from ChatGPT, Grok and Gemini. It says a Grok-built narrative implicated an Iranian citizen, who was arrested and cleared; a defence lawyer calls the indictment "outrageous".

Baltimore sues X, xAI and SpaceX over Grok sexual deepfakes

Who did it: HYBRID DiCello Levitt

Baltimore's mayor and city council have sued X Corp., xAI and SpaceX in the Circuit Court for Baltimore City, alleging Grok produces and spreads non-consensual sexualised images, including of minors, in breach of the city's Consumer Protection Ordinance. The complaint cites estimates that millions of such images were generated in days, thousands depicting minors.

Music publishers ask judge to reject Anthropic's fair use defence

Who did it: HYBRID Reuters

Universal Music Group, Concord and ABKCO asked a federal judge in San Jose to rule before trial that Anthropic infringed their copyrights by copying song lyrics to train Claude, and to reject its fair use defence. The publishers, who sued in 2023 over lyrics from at least 500 songs, say Claude reproduces lyrics on demand; Anthropic denies the allegations.

Tesla faces Australian legal fights over self-driving hardware limits

Who did it: HUMAN The Sydney Morning Herald

Tesla faces a NSW tribunal claim and a Federal Court class action alleging it sold Australian owners Full Self-Driving, priced at $10,100, for older HW3 cars incapable of fully running it. A tribunal member rejected Tesla's bid to dismiss or stay one owner's misleading-conduct claim, letting it proceed alongside the class action.

Oregon appeals court fines lawyer record $10,000 over AI fakes

Who did it: HYBRID The Oregonian

The Oregon Court of Appeals fined Salem attorney Bill Ghiorso $10,000, a record for an Oregon appellate attorney, for a brief with 15 bogus citations and nine quotes 'contrived from thin air'. Ghiorso said his paralegal's Google research apparently surfaced AI-generated content; the court capped the fine, which could have reached $16,500, citing his recent medical trouble.

Alkistis Protopsalti reports AI fake ad video to cybercrime police

Who did it: HUMAN iefimerida

Greek singer Alkistis Protopsalti went to the police cybercrime unit with her lawyers after an AI-generated video on Facebook showed her advertising various products, made by people seeking to take money from consumers. She posted a warning about the 'fake video' of her advertising 'rubbish', demanding it be taken down and calling those behind it 'idiots'.

Mantova court fines claimant over misattributed Cassazione rulings

Who did it: HUMAN AI Hallucination Cases database

The Tribunale di Mantova rejected a claimant's case after finding that eight Cassazione rulings it cited contained nothing supporting its argument, the judge suggesting this could only be explained by unverified use of AI. The claimant was ordered to pay €9,142 in costs plus €2,250 to the defendant and €2,250 to the Cassa delle ammende.

CanLII settles copyright suit against AI legal assistant Caseway

Who did it: HUMAN Canadian Lawyer

The Canadian Legal Information Institute has settled, on confidential terms, the copyright suit it filed in November 2024 in the Supreme Court of British Columbia accusing Caseway AI of stealing content from its legal database, over 120 gigabytes and 3.5 million records, and putting it behind a paywall. Caseway called it a 'forward-looking resolution' and a first in legal tech.