AI Misconduct.com

The daily rap sheet of the machine age.

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

Wednesday, August 13, 2025

7 items

WA Police warn of Veo deepfakes misrepresenting firearm laws

Who did it: HUMAN The West Australian

WA Police and Consumer Protection WA issued a warning about two AI-generated videos made with Veo and posted to a Facebook group. One shows an officer refusing a firearms licence to ADF veterans; the other shows police taking a gun from a crying farmer. Asked about them, the page's admin replied with a Veo video of police arresting a gorilla.

Google Lens AI overviews repeated debunked claims about viral clips

Who did it: AI Full Fact

A Full Fact investigation found Google Lens's AI overviews failed to recognise inauthentic content or otherwise shared false claims for at least ten widely shared clips, including placing footage that appears to show a beach in Goa on Dover Beach, as in a false asylum-seeker claim. Google said errors it could reproduce came from visual search results, not the overviews.

Study finds AI browser assistants harvesting health and banking data

Who did it: HYBRID EurekAlert!

A UCL and Mediterranea University of Reggio Calabria study of nine generative AI browser extensions found several sent full webpage content to their servers; Merlin captured form inputs such as banking or health data, and Sider and TinaMind shared queries and IP addresses with Google Analytics. Researchers say some violated US laws including HIPAA and FERPA.

Columbia, Missouri man charged over allegedly AI-generated abuse images

Who did it: HUMAN KOMU 8

Charles Hooton, 63, of Columbia, Missouri, was charged Tuesday with felony possession of child pornography after police say they found 135 child sexual abuse images on his phone, 130 of them believed AI-generated. According to the probable cause statement, he admitted using an AI app to generate sexual images of minors; he is held without bond.

Florida appeals court admonishes wife for 'counterfeit' brief, skips sanctions

Who did it: HYBRID AI Hallucination Cases database

A Florida appellate court admonished a self-represented wife for a "counterfeit brief" containing two hallucinated case citations, according to the AI Hallucination Cases database, which lists AI use as implied. The court declined sanctions because she was defending a judgment unrepresented, the husband sought none and she had no other frivolous filings, but warned it would be less lenient next time.

GAO warns Oready a second time over fake protest citations

Who did it: HYBRID AI Hallucination Cases database

GAO dismissed Oready LLC's protest over an Interior Department school counselor solicitation as academic on August 13, 2025, and warned the firm for a second time about fictitious or misleading citations. It rejected Oready's blaming of "manual mismatches in secondary summaries," saying the errors bear the hallmarks of unverified AI or large-language-model use.