AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 3, 2026
4,329 items in the archive
RSS feed

Tuesday, February 18, 2025

7 items

AP finds Israeli military leaning heavily on US-made AI

Who did it: HYBRID Associated Press

An AP investigation found that the Israeli military's use of Microsoft and OpenAI artificial intelligence spiked last March to nearly 200 times its level before the October 7 attack. Officers described targeting errors caused by faulty data and mistranslations from Arabic. The military says humans always review targets, and OpenAI says it has no partnership with Israel's military.

South Korea regulator says DeepSeek was communicating with ByteDance

Who did it: HUMAN BBC

South Korea's Personal Information Protection Commission told Yonhap it "confirmed DeepSeek communicating with ByteDance", though it has yet to confirm what data was transferred or to what extent. DeepSeek, downloaded over a million times in the country, had been pulled from Apple and Google's app stores, and the regulator says the company acknowledged failing to take South Korean privacy laws into account.

Report: fake FTX claim seller may have used AI face-swap

Who did it: HUMAN CoinDesk

At least two unnamed firms buying FTX bankruptcy claims lost at least $5.6 million to someone posing as a claim seller, according to an investigation by Inca Digital, which concluded the fraudster likely used AI deep-faked video in calls. Buyers were also shown faked identification and real claim data, and the funds were laundered through exchanges including Binance, the report says.

Suit alleges TikTok geotargeted railroad suicide videos to a teen

Who did it: HYBRID New York Post

The parents of a 16-year-old who died by suicide in 2022 allege TikTok's algorithms bombarded him with thousands of suicide-related videos and used his location data to send railroad-themed ones, as he lived near train tracks. Their Feb. 5 filing opposes TikTok's bid to dismiss the suit, which argued app content is protected speech.

Jaclyn Victor says she'll go to police over fake nude video

Who did it: HUMAN Malay Mail

Singer Jaclyn Victor posted a screenshot of a doctored, likely AI-generated video showing her naked and being arrested, saying she was going to the police. "Jaclyn Victor ditangkap" (Jaclyn Victor arrested) has become a suggested Google search for her name; the video cites a fake news article that does not exist, and the face editing is poorly done.

Nevada judge tosses suit citing nonexistent cases, skips sanctions

Who did it: HYBRID AI Hallucination Cases database

A Nevada federal judge dismissed a pro se civil rights complaint with prejudice on Feb. 18, 2025, for violating pleading rules and citing non-existent cases such as "Spokane v. Douglass". Judge Cristina D. Silva wrote that either the plaintiff fabricated the cases or AI invented them, followed by lying about their origin, and declined to pursue sanctions.

Robert Tsao sues pundit over allegedly faked affair photos

Who did it: HUMAN The Epoch Times

Commentator Xie Hanbing claimed UMC founder Robert Tsao, lead petitioner in the recall of KMT legislator Hsu Chiao-hsin, had an extramarital affair and published several photos. Tsao says the intimate photos are composites, and on Feb. 18 announced criminal and civil suits against Xie for forgery and defamation, seeking 100 million in damages to donate to the recall campaign.