A Delaware Superior Court judge on July 24 rejected Google's bid to escape conservative activist Robby Starbuck's suit claiming its Bard chatbot defamed him, allegedly by spreading false sexual assault accusations and linking him to white supremacist Richard Spencer. Judge Meghan A. Adams also declined to disregard, before discovery, his theory that the chatbot can understand its own programming.
An Arkansas family has sued xAI after Bentonville police said photographer Russell Bloodworth III used Grok to turn photos of clients, including their daughter, into child sexual abuse material. Bloodworth was arrested on June 10 and charged with 200 counts. The suit seeks damages, deletion of the images and safeguards; the family's lawyer says more suits will follow.
A threat actor ran the open-source Hermes AI agent in unattended "YOLO" mode to automate post-exploitation work in an alleged breach of Thailand's Ministry of Finance, according to threat intelligence firm Hunt.io and a security researcher, who found exposed logs, web shells and stolen credentials. The ministry has not confirmed a breach; the logs show a human operator directing the agent.
Four Twin Cities men pleaded guilty to wire fraud over about $2.2 million taken from a Minnesota Medicaid housing program, prosecutors announced July 24. Prosecutors say they billed for roughly 350 clients while providing little or no help, and when asked for documentation "fabricated records using ChatGPT," according to the DOJ. Each faces up to 20 years.
Queensland police have charged a 52-year-old Cleveland man with 43 offences, alleging he used AI to put at least eight women's faces onto explicit images and displayed them, with names and employment details, on posters advertising sex across southeast Queensland since 2021. Police also seized firearms in Wednesday's raid; he is remanded in custody.
Robert John Thomas, 73, a former teacher at Gold Coast school A.B. Paterson College, has been charged with making and possessing child exploitation material after police alleged he used his work laptop and AI software to generate explicit images of children and adults. The school says it fired him and notified police; he appears in court August 6.
A Tacoma federal judge imposed a Rule 11 sanction on plaintiff's lawyer Jocelyn Stewart on July 24 after dozens of false citations and fictitious quotations appeared across at least five filings in LeDoux v. Outliers. Stewart declared that ChatGPT and Claude generated them and she "did not verify the output"; her "corrected" motion swapped in real quotes that changed the meaning, the court found.
Brazilian beauty influencer Karen Bachini, who in June launched a R$149 monthly scrapbooking subscription selling illustrations presented as her own, admitted using AI to "finish" the drawings after artists and other influencers spotted telltale signs in buyers' photos. In an apology video she called it "a very ugly thing", said there was no excuse, and cited past insecurity.
A Marrero woman first treated as a victim in an online threats and extortion case was arrested after platform subpoenas pointed back to her, the Livingston Parish Sheriff's Office said. Sheriff Jason Ard said an interview established that Haydee Ortiz, 29, herself made the fake accounts, AI-generated nude images, rape threats and extortion attempts. She was released on $150,000 bond.
A federal grand jury indicted Daniel Bostwick, 69, of Hillsboro, a registered sex offender on parole with device monitoring, for possessing obscene visual representations of child sex abuse after authorities say they found dozens of apparently AI-generated images on his phone and a thumb drive. He pleaded not guilty; a jury trial is set for September 22.
