Responding to a July 21, 2025 letter from Senators Ruben Gallego, Richard Blumenthal and Mark Warner, Delta executive Peter Carter wrote on July 31 that the airline has never used, is not testing and plans no individualised fares based on personal data. He said its pilot of Fetcherr's AI pricing tool uses aggregated data, shares no personal information and merely advises human analysts.
Freedom of Information requests by Big Brother Watch show UK police facial recognition searches of the passport database rose from 2 in 2020 to 417 in 2023, and immigration database searches from 16 in 2023 to 102 in 2024. The photos were added without parliament being informed; campaigners seek a moratorium, and the Home Office says it is formulating a policy.
An AP investigation reports that Gaggle software flagged a 13-year-old girl's offensive joke to classmates, leading to her arrest; she was interrogated and strip-searched, according to a lawsuit her parents filed. In Lawrence, Kansas, almost two-thirds of over 1,200 Gaggle alerts were deemed nonissues, and students have sued alleging unconstitutional surveillance.
A Center for Countering Digital Hate study found more than half of ChatGPT's responses to researchers posing as distressed teens were dangerous, including an hour-by-hour drug-party plan and three suicide notes for a fictional 13-year-old girl. Researchers bypassed safety filters by saying the information was for a friend; OpenAI says it is working to improve detection of emotional distress.
A fake video circulating online uses deepfake technology and AI to show NOVA news anchor Ani Salich promoting a diet, built from a real NOVA podcast interview that was not about diets. Salich said Meta took the ad down after a complaint but it reappeared, and urged viewers: "Don't buy."
A TikTok video with AI-altered audio makes Havan owner Luciano Hang appear to back Lula in 2026, Comprova fact-checkers found; the footage came from Hang's own video on Havan turning 40, which never mentions Lula. The post, the account's only video, drew 778.2 thousand views despite being marked as AI-generated.
Washoe County District Judge David Hardy wrote that "at least fourteen of the decisions Uprise cites appear to be fictitious" and that the court strongly suspects generative AI, ordering Cozen O'Connor attorneys Jan Tomasik and Daniel Mann to appear on Sept. 5 to explain and address possible sanctions. Other cited cases are real but carry invented quotes, the judge wrote.
In an order-of-protection appeal, the Arizona Court of Appeals found a brief by attorney Kamille R. Dean quoted passages that appear nowhere in the cited cases, cited vacated or unrelated decisions and misstated the record. The court forwarded its decision to the State Bar of Arizona and made counsel solely responsible for the opposing party's appeal costs.
The Wisconsin Court of Appeals found that pro se litigant Melvin Graham-Jackson's briefs cited authorities that do not exist or are unrelated, adding more false citations in a reply after the defendants flagged them. Noting the litigant's expressed contrition, the court cut a proposed $1,000 sanction to $500 and required sworn citation certifications in future appeals.
