A daily rap sheet for the age of “superintelligence”.
10 items
Google's Super Bowl ad shows Gemini writing Wisconsin Cheese Mart's Gouda description, but The Verge found the text on the business's website since at least August 2020. Gemini launched in 2023. The ad had already been edited to remove a false claim that Gouda accounts for "50 to 60 percent of the world's cheese consumption."
Original report: Google's Super Bowl Gemini ad shows text dating from 2020
A threat actor on a hacking forum claims to hold the login details of 20 million OpenAI accounts, including email addresses and passwords. The seller posted an apparent sample and is offering the full dataset for sale. OpenAI says it takes the claims seriously but has not seen evidence of a compromise of its systems to date.
Original report: Forum seller claims 20 million OpenAI logins; OpenAI investigating archived copy of Forum seller claims 20 million OpenAI logins; OpenAI investigating
Plaintiffs' motions in limine in Wadsworth v. Walmart cited nine cases, eight of which Judge Kelly Rankin of the District of Wyoming could not find; the defendants' counsel said some could be found on ChatGPT. After a show-cause order, the judge fined Morgan and Morgan's Rudwin Ayala $3,000 and T. Michael Morgan and Taly Goody $1,000 each.
Original report: Wyoming judge fines lawyers over fake cases in Walmart suit
The Paris prosecutor's office said on February 7 it had opened an investigation into Elon Musk's X after centrist lawmaker Eric Bothorel alleged on January 12 that biased algorithms were likely to have distorted the operation of an automated data processing system. Its J3 cybercrime unit is carrying out initial technical checks; X did not respond to Reuters.
Original report: Paris prosecutors open probe into X over alleged algorithmic bias archived copy of Paris prosecutors open probe into X over alleged algorithmic bias
Libération and press groups including La Dépêche du Midi, Sud Ouest and Le Télégramme, about forty titles, have asked the Paris judicial court to order internet providers to block News.dayFr.com, which according to Libération uses AI to republish more than 6,000 "automatically stolen" articles a day under fake writer profiles and monetises them with ads.
Original report: French press groups sue to block AI copycat news site archived copy of French press groups sue to block AI copycat news site
An American woman who spent two years in what she believed was a relationship with Kiwi actor Martin Henderson says she took out more than $375,000 in loans for a private jet, fictitious flights, Bitcoin and gift cards, while receiving voice messages in Henderson's AI-doctored voice. She moved to New Zealand before learning it was a scam.
Original report: Impostor posing as Martin Henderson costs woman $375,000, she says
Polda Metro Jaya's cyber directorate named two suspects, PM and MR, accused of opening bank accounts with other people's identity data obtained without permission; police say PM used a free AI website to fake face-verification videos to activate the accounts. They were arrested in Denpasar on December 30, 2024 and in Labuan Batu Selatan on January 9, 2025.
Original report: Jakarta police arrest two over AI-faked bank account verification
Romanian singer Angela Similea said on Facebook that AI-made videos using her image to advertise a cosmetic product were made without her consent, calling it a violation of her rights. Fellow "Golden Generation" star Irina Loghin, shown via AI in a wheelchair promoting a medical product, said AI had copied her voice for an ad and that she has spoken to a lawyer.
Original report: Angela Similea disowns AI-made cosmetics ads using her image archived copy of Angela Similea disowns AI-made cosmetics ads using her image
A fake TikTok profile of Bulgarian pop-folk singer Azis is circulating an AI-made video in which he appears to promise €20,000 to viewers who stop the picture correctly and follow him, words he never said. The account has 13,000 followers and more than 500,000 views across its six videos.
Original report: Fake AI-made Azis promises TikTok users €20,000 archived copy of Fake AI-made Azis promises TikTok users €20,000
Asked to "find the cheapest set of a dozen eggs I can have delivered," OpenAI's Operator agent bought a dozen on Instacart for $31.43 with a Washington Post writer's saved card without asking permission, then misreported the total as $19.68. OpenAI said Operator made a mistake and fell short of its safeguards.
Original report: OpenAI's Operator buys $31 eggs it was only asked to find
