Two secondary schools run by the University of Buenos Aires say they are investigating reports that students made AI-manipulated sexualised images of female classmates, which were allegedly sold. At the Colegio Nacional, sources told LA NACION, eight pupils aged 13 and 14 allegedly shared such images in a WhatsApp group. They were suspended as a precaution, and the Pellegrini says expulsion is possible.
Traci Tamiko Eto, Mayo Clinic's former director of research operations, has sued in federal court in Minnesota. The suit alleges she was sidelined and fired after objecting to lax oversight of AI research, and that Mayo leaders pressured review boards and researchers tried to bury results showing a 67% error rate in its MAYA assistant. Mayo said it does not comment on active litigation.
NHTSA Administrator Jonathan Morrison told autonomous vehicle developers the agency had found "a clear pattern" of driverless cars entering emergency scenes, blocking ambulances and firefighters, and failing to recognise flares and flashing lights. He demanded solutions by month's end. The letter names no company. TechCrunch previously found at least six incidents in which first responders had to move Waymo vehicles.
Meta has built its Muse Image generator into Instagram, letting users make images from photos on public profiles simply by tagging a username. Public accounts are opted in automatically, and Meta does not notify people when their photos are used. Switching the setting off does not remove images that have already been generated.
Gangseo Police Station said it is investigating a man who allegedly covered the recording indicator light on his Meta AI glasses. He is accused of secretly filming a woman he was on a date with and posting the footage on social media. Police said they will examine his digital devices and intend to pursue sex-crime charges.
Bradford magistrates convicted landlord Mindaugas Palaima, in his absence, of offences including running an unlicensed HMO and failing to fit smoke alarms. Prosecutors said that after joking he could send an AI-generated image, he sent a photo of a ceiling smoke detector, and officers who visited the next day found none. He must pay £11,218.
Berlin think tank Agora Digitale Transformation put 675 news queries to AI systems over five working days in May. It found that Anthropic's Claude drew seven times on domains of the "Pravda network", which experts describe as a Russian influence operation, and the study says Claude presented them on a par with journalism. Anthropic said it is working to train Claude to be politically balanced.
A writer for The Drive says Plymouth, Minnesota, police tracked their test Range Rover for days with Flock cameras, then boxed it in at a Kohl's lot with hands on their guns. According to the account, a stolen plate had been entered as "34 DTM", and Flock's system ignored the small middle digits on New Jersey manufacturer plates.
Ahead of Typhoon Bavi, some Chinese bloggers posted their own track and rainfall forecasts built from open-source AI weather models, some charging about 10 yuan per local forecast. Legal academics told China Media Group reporters this may breach the Meteorological Law, which reserves public forecasts to official weather stations and allows fines of up to 50,000 yuan.
Magistrate Judge Bill Davis of the Eastern District of Texas recommended granting staffing firm Ampcus's motion to dismiss a pro se Fair Credit Reporting Act suit. A Westlaw editor's note says the ruling discusses unverified citations likely generated by AI tools used by a party or counsel, which have been preserved in the official record.
Florida's Second District Court of Appeal affirmed without comment but wrote to caution a pro se appellant against using generative AI without verifying it. The court said the appellant's briefs were "replete with mismatches" between legal propositions, case names and reporter citations, and warned that such failures expose filers to potential sanctions.
