A legal filing published Tuesday shows Anthropic has reached a preliminary settlement in the authors' class action Bartz v. Anthropic. Judge William Alsup had ruled that training on the books was fair use but that downloading them from shadow libraries such as LibGen was piracy. That ruling left Anthropic facing a December trial and potential penalties amounting to billions.
The parents of a 16-year-old California boy who died by suicide in April 2025 have sued OpenAI and Sam Altman in San Francisco. Their complaint alleges ChatGPT discussed suicide methods with him for months, offered to draft his suicide note and analysed his noose hours before his death, and that OpenAI rushed GPT-4o to market despite safety warnings.
An NYPD facial recognition search flagged a man as a "possible match" for a flasher described as about 5 feet 6 and 160 pounds; he is 6-foot-2 and 230 pounds, and his phone data put him about 12 miles away. After the victim picked him from a photo lineup, he spent more than two days in jail before the case was dismissed in July.
Dispatch's founder says a pitch from writer 'Margaux Blanchard' about Gravemont, a secret Colorado forensics training town, described a place that does not exist, and neither, it emerged, did she. Wired had already pulled her Minecraft-weddings story over fabricated sources, and at least six publications have since pulled her articles. The founder casts her as one of many submitting AI pitches.
Who did it: HYBRID AI Hallucination Cases database
A federal magistrate judge in Oklahoma struck Chapter Kris Jackson's amended motion to disqualify a lawyer for BOK Financial entities. The judge found she had only removed quotation marks from fake quotations flagged earlier and left other fabricated citations. On September 29, 2025, the court ordered her to show cause under Rule 11 and set a sanctions hearing for October 21.
Actor Tay Ping Hui warned fans on Instagram on Aug 22 that his image had been used in fake AI marketing videos promoting 'medication', and urged them not to click links or buy the product. He called the scammers lazy, noting the AI version of him 'looked nothing' like him; the video also used real footage from Allan Wu's podcast.
Forty-four attorneys general sent an open letter to 11 chatbot and social media companies, including Meta, OpenAI, Google and xAI, warning they will "answer for it" if they knowingly harm children. The letter cites reporting on internal Meta policies stating "It is acceptable to engage a child in conversations that are romantic or sensual." Replika called such conduct indefensible.
Police in Kankakee and Bradley are investigating calls placed to non-emergency lines at about 4:33 a.m. Monday threatening an unspecified Kankakee school and Bradley East Elementary School. Police believe the calls were AI-generated robocalls, in which the caller claimed to be "forced to make the threats," and deemed them not credible, though police presence at schools was stepped up.
