According to correspondence obtained by Reuters, Tesla gave regulators in the Netherlands and Sweden self-published FSD safety statistics that independent researchers call misleading marketing. A presentation sent to Swedish regulators after the Dutch approval on April 10 claimed FSD could have saved 32,000 lives, a figure researchers say assumes every U.S. vehicle, motorcycles included, became a Tesla; Tesla did not respond.
KPMG has taken down a report on the future of AI in which, according to GPTZero, 40 of 45 citations appeared at least partially hallucinated, including a claim that JR East uses agentic AI sourced to a 2019 press release. GPTZero alleges an LLM research tool's output was copied in without verification; KPMG says it is reviewing the circumstances.
The DOJ and Homeland Security Investigations seized CFAKE.com and SOCFAKE.com, which allegedly hosted nonconsensual AI-generated sexual images of famous women including politicians, first ladies and royalty, in what appears to be the first publicly announced domain seizure under the TAKE IT DOWN Act. The case began with a tip from Italy's Postal Police; French investigators arrested a suspect in Nice on June 10.
A proposed federal class action filed on behalf of a customer alleges usage limits on Anthropic's plans costing $100 and $200 a month, which promised five and 20 times the usage of its Claude Pro tier, were difficult to understand and changed without clear notice, the Wall Street Journal reported. The suit points to a gap between what customers expect and what they receive.
US District Judge Eumi Lee on June 11 denied Meta's motion to dismiss Strike 3 Holdings' copyright claims, which allege Meta IP addresses torrented at least 2,396 of its adult films 6,008 times between 2018 and 2025 to train AI. Meta calls the claims bogus, saying it takes deliberate steps to avoid training on such material.
In sealed civil and criminal proceedings, Unico accuses Serasa Experian of accessing its facial-biometric identity checks via Skill Tecnologia, a firm authorised only for Banco do Brasil, according to people who spoke to Folha; an expert report it commissioned reportedly found at least 1.4 million irregular transactions. Serasa, reportedly the target of a search warrant executed in São Paulo, denies the accusation.
Shanghai transit police say a 50-year-old woman from outside the city, after seeing reports of a little wall plaster falling at Line 18's Jiangpu Road station on June 12, used an AI tool the next day to generate a video of a large-scale ceiling collapse to attract traffic, posting it to her WeChat Channels account. She has been placed in administrative detention.
Judge Robert S. Lasnik struck the motion to dismiss filed by Kathy Q. Hao, a licensed attorney representing herself against Amazon's trademark-fraud suit, after Amazon said it cited nonexistent cases; Hao acknowledged relying partly on AI-assisted research. The judge declined monetary sanctions for now, allowed refiling with personally verified authorities, and warned that fabricated citations next time may bring sanctions.
Greek journalist Katerina Zarifi told MEGA's "Buongiorno" that AI-generated images of her in a wedding dress, with a false report that she is marrying, brought congratulations and offers to make the dress, the third time, she says, she has been an AI victim. A hearing in an earlier slimming-pills case was postponed; she says the person who started it has been referred to trial.
