AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 3, 2026
4,238 items in the archive
RSS feed

Tuesday, March 25, 2025

7 items

Meta tells court training Llama on authors' books was fair use

Who did it: HUMAN Reuters

Meta Platforms asked a federal judge in San Francisco on Monday to rule that using books by Ta-Nehisi Coates, Sarah Silverman and others to train its Llama model was fair use, and to throw out the authors' 2023 lawsuit. The authors allege Meta fed pirated copies into its models without paying; Meta says Llama does not replicate the books or substitute for reading them.

Chinese livestream pirates face-swap hosts with AI to sell knockoffs

Who did it: HUMAN 法治日报

According to Legal Daily, sellers on Chinese short-video and social platforms record other merchants' livestreams, swap the hosts' faces with AI and replay them, many to sell knockoffs. In two hours a reporter found four face-swapped streams selling the same shirt at different prices, and was kicked out of one after asking. A complaint site lists over 3,000 related complaints.

Pro se plaintiff's AI-sourced fake cases earn warning, not fine

Who did it: HYBRID AI Hallucination Cases database

Vladimir Kruglyak, suing Home Depot without a lawyer, admitted a reply brief cited fictitious cases and misstated holdings, saying he had relied on free generative AI tools. On March 25, 2025 a Virginia federal magistrate judge found him negligent but not in bad faith. The judge imposed no monetary sanctions, ordered him to flag AI-researched cases, and denied his request to amend.

AI nude image victim backs Florida bill closing creator loophole

Who did it: HUMAN WEAR

A victim whose photo was edited into AI-generated child sexual abuse imagery urged a Florida Senate committee on Tuesday to back Senator Don Gaetz's Senate Bill 1180. The male student who made images of several students was not arrested because he did not distribute them. A girl who sent them out got probation after pleading no contest.

North Bay police warn of deepfake Musk crypto pitches

Who did it: HUMAN CTVNews

The North Bay Police Service said Tuesday it has received several reports of AI-generated videos of Elon Musk promoting cryptocurrency investments circulating online. A spokesperson said the scams began with a link and video on social media, followed by further contact with potential victims. The spokesperson warned that other prominent people's likenesses may also be used.