A consumer refused a mobile phone contract over Dun & Bradstreet Austria's automated credit assessment won an Austrian regulator's order for an explanation, but the firm withheld some information as trade secrets. On 27 February 2025 the CJEU ruled controllers must explain "the procedure and principles actually applied" and cannot use trade secrets to justify a blanket refusal.
The US Department of Labor has been investigating data-labelling startup Scale AI, valued at $13.8 billion, for compliance with the Fair Labor Standards Act since at least August 2024, TechCrunch reports. Scale AI says regulators under the previous administration misunderstood its work and talks have been productive; the department would neither confirm nor deny any investigation.
Israel's Unit 8200 is building a ChatGPT-like language model trained on millions of Arabic conversations obtained through surveillance of Palestinians in the occupied territories, according to an investigation by +972 Magazine, Local Call and the Guardian. One source said it was trained on 100 billion words; the IDF Spokesperson said tool use follows "a rigorous process led by professionals."
Secretary of State Marco Rubio is launching an AI-fueled "Catch and Revoke" effort to cancel visas of foreign nationals who appear to support Hamas, senior State Department officials told Axios, including AI-assisted reviews of tens of thousands of student visa holders' social media accounts. Abed Ayoub of the American-Arab Anti-Discrimination Committee called it a free speech issue relying on "faulty technology."
According to a YouTube alert, scammers are sending creators private videos in which an AI-generated Neal Mohan announces YouTube Partner Program changes, with links to phishing sites that steal Google credentials. Hijacked channels are often used for cryptocurrency scam streams. YouTube says it never contacts users through private videos, a point the phishing emails themselves helpfully mention.
Four more French families have filed complaints against TikTok, accusing it of failing to moderate content on suicide, self-harm and eating disorders and of not keeping minors away from it; two of the families lost teenage daughters to suicide. Seven families had already filed in November 2024, and that case is ongoing, said lawyer Laure Boutron-Marmion.
Since late 2024, AI avatars made with the Captions app have flooded TikTok claiming "13 years as a coochie doctor" and offering suspect advice such as using lemon balm in place of Ozempic, the Daily Dot reports. Dozens of videos share the same hook, and the same accounts also claim expertise in guts, plastic surgery and dietetics.
A search for "mein kampf positive reviews" returned a Google featured snippet drawn from Amazon's AI summary of customer reviews, which says readers describe the book as "a true work of art" and appreciate its "insightful and intelligent rants." Alexios Mantzarlis of Cornell Tech spotted it; neither company immediately responded to requests for comment.
Spain's Partido Popular posted an AI-made video comparing the "Koldo case" to the TV show La Isla de las Tentaciones, set in the Dominican Republic and billed as "La Isla de las Corrupciones." The Dominican Republic issued a statement of "most forceful rejection," and Prime Minister Pedro Sánchez apologised in Spain's name, saying he was "extremely ashamed."
Testing eight AI search tools on 1,600 queries, the Tow Center found they misidentified news articles in more than 60 percent of cases, with Perplexity wrong on 37 percent and Grok 3 on 94 percent; Grok 3's citations led to error pages 154 times out of 200. Several tools also appeared to bypass publishers' robots.txt blocks.
A video shared by a Facebook page overlays AI-generated audio on footage of Finance Minister Nirmala Sitharaman from the December 21, 2024 GST Council press conference, having her promise that Rs 21,000 can become more than half a million rupees in a month, BOOM found. Its link led to a page impersonating The Times of India.
A Nord University student has been barred from all Norwegian universities and colleges through 2026 after a board found AI was used to generate parts of a submission, Khrono reports. The student, previously caught cheating, admitted only to using AI to find sources and assess the text. Khrono says it is the first penalty this severe.
Weber County deputies arrested Sean Patrick Layton, 46, of West Haven, for investigation of 20 counts of sexual exploitation of a minor after a cybertip from a photo editing app. According to the sheriff's office, he admitted creating child sexual abuse images using AI, including one bearing a local girl's face taken from social media.
