AI Misconduct.com

The daily rap sheet of the machine age.

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

Monday, May 12, 2025

10 items

Veritone pitches clothing-based tracking AI as facial recognition workaround

Who did it: HUMAN MIT Technology Review

Veritone's Track tool, used by 400 customers including US police departments and universities, identifies people by body size, gender, hair, clothing and accessories rather than faces; CEO Ryan Steelberg says it permits tracking where facial recognition isn't legally allowed. An ACLU analyst called it "a potentially authoritarian technology", though it is unclear whether it is used where facial recognition is banned.

Kids' homework chatbots gave Forbes fentanyl and date-rape drug recipes

Who did it: AI Forbes

In Forbes tests, Knowunity's SchoolGPT gave step-by-step fentanyl synthesis instructions once told it inhabited an alternate reality, and suggested 967 calories a day for a hypothetical teen dieter; a CourseHero homework bot explained how to synthesize the date-rape drug flunitrazepam. Knowunity CEO Benedict Kurz said the company was working to exclude such responses, which stopped after its tweaks.

Court restrains Craig Wright's lawsuits, noting ChatGPT-style fake citations

Who did it: HYBRID AI Hallucination Cases database

On 7 March 2025 Mr Justice Mellor granted COPA and SquareUp a three-year General Civil Restraint Order against Craig Wright, referred his conduct to the Attorney General and ordered him to pay £100,000 costs. Reasons handed down on 12 May cite his appeal papers, filed while representing himself, as "bearing signs of creation by ChatGPT" and citing non-existent authorities.

German consumer group threatens legal action over Meta AI training

Who did it: HUMAN TechRadar

Verbraucherzentrale North Rhine-Westphalia sent Meta a cease and desist letter on April 30, 2025, demanding Facebook and Instagram drop plans to train AI on EU users' public posts and interactions from May 27, and threatened further legal action. Meta says its approach complies with EU law; noyb's Max Schrems backed the group's action.

YouTube cuts ads on more AI fake trailer channels

Who did it: HUMAN Deadline

Following a Deadline investigation, YouTube suspended ad revenue on Screen Trailers and Royal Trailer, alternative accounts run by the creators behind Screen Culture and KH Studio, whose AI-heavy concept trailers in some cases fool viewers into thinking they are authentic. The main channels were suspended in March; Deadline found Warner Bros. Discovery and Sony had claimed ad revenue on Screen Culture trailers.

NRLW star Jaime Chapman decries repeated AI-doctored photos of her

Who did it: HUMAN Mail Online

Gold Coast Titans and NSW player Jaime Chapman said on Instagram that someone had again used AI to doctor a photo of her into a fake image in a revealing outfit. "This has happened a few times now and it needs to stop," she wrote, calling the effect "scary" and "damaging".

Deepfaked Cypriot politicians front investment scam that threatens victims

Who did it: HUMAN philenews.com

The Cyprus Consumers Association says investment scams use AI videos of figures including the President, Averof Neofytou and Haris Georgiades recommending investments, and that victims who discover the deception and demand explanations receive threats. Pitches typically start at €250; one case exceeded €40,000 in losses, and 14 people reported their experiences in a single day, its president Marios Drousiotis said.

Mississippi judge finds pro se plaintiff's cited cases don't exist

Who did it: HUMAN AI Hallucination Cases database

Staying a suit pending an officer's appeal on 12 May 2025, Judge Michael P. Mills said the court's earlier ruling sustaining the officer's qualified immunity defense was largely a sanction on a pro se plaintiff for citing cases that do not exist, and called her suggestion of an innocent mistake "quite implausible".

Deepfake Fauci and AI doctors hawk sexual cures on TikTok

Who did it: HYBRID AFP

AI-generated videos on TikTok are peddling unproven sexual remedies, including "AI doctor" avatars and a deepfake of Anthony Fauci promoting a prostate supplement, AFP reports; some use carrots as a euphemism, apparently to evade moderation. Researcher Alexios Mantzarlis found hundreds of YouTube ads last year using celebrity deepfakes to promote erectile dysfunction cures.