AI Misconduct.com

The daily rap sheet of the machine age.

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

Friday, May 30, 2025

8 items

Impersonator, possibly AI-voiced, poses as Susie Wiles; FBI investigates

Who did it: HUMAN The Guardian

The FBI is investigating an impersonator of White House chief of staff Susie Wiles who texted and called her contacts, including prominent Republicans, asking one lawmaker for a pardon list and, in another case, for a cash transfer, the Wall Street Journal reports. Contacts said some calls came in a voice that sounded like Wiles, leading some to believe AI was mimicking it.

Delhi High Court orders takedown of AI fakes of Sadhguru

Who did it: HUMAN India Today

In a personality rights case filed by Sadhguru and Isha Foundation, the Delhi High Court issued an interim order directing online platforms to take down fake AI-generated content misusing the spiritual leader's name, image and personality. According to the foundation, the scams include morphed images of a fictitious arrest and misleading financial investment advertisements.

Houston FBI analyst charged over abuse images; cache included AI fakes

Who did it: HUMAN KPRC

Brian Vincent Rausch, a Houston FBI analyst, faces four felony charges in Harris County, including two counts of possessing child pornography involving 500 or more images, according to court records. Investigators say the files also held AI-generated images nearly impossible to distinguish from real child victims. Rausch posted bond, with a court appearance set for May 30.

Tesla and Dutch regulator accused of loophole in Autopilot approval

Who did it: HUMAN Zembla

Dutch regulator RDW approved Tesla's Autopilot for Europe by reading the rules as allowing a chain of single manoeuvres, a "loophole" according to a Swedish official quoted by Zembla; an RDW whistleblower alerted other countries. The RDW denies using a loophole and says the system met requirements. Tesla did not respond.

Paralegal's AI-assisted factum cites fake cases; client ordered to pay costs

Who did it: HYBRID AI Hallucination Cases database

At an Ottawa Small Claims Court hearing, paralegal Nima Kiani acknowledged a factum was prepared with AI assistance, then supplied "revised" citations for three nonexistent cases that the court found were not the cases originally cited, and admitted a fourth does not exist. The plaintiff was ordered to pay $1,660.30 in costs; the judge would have ordered the representative to pay had the rules allowed.

Deepfake pop stars sing praises of Burkina Faso's junta leader

Who did it: HUMAN France 24

Journalists counted at least 294 AI-generated music videos on at least 127 YouTube channels, with more than 8.6 million views, showing deepfaked stars including Beyoncé, Rihanna and R. Kelly singing "God Protect Ibrahim Traoré". A Nigeria-based creator claimed to be behind most of the popular ones; a researcher calls it a coordinated campaign but lacks proof the government is behind it.

AI clickbait campaign targets Rockingham County students, sheriff warns

Who did it: HUMAN Facebook

The Rockingham County Sheriff's Office warned of an online campaign using AI-created clickbait, including a student's likeness, to lure students and residents to websites with harmful pop-ups or malicious attacks. According to the office, Proxyware found nearly 100 more sites sharing the same tracking code, pointing to a centralized operation based in South Africa.