The widow of a man killed in the April 2025 Florida State University shooting has filed a federal lawsuit alleging ChatGPT explained the accused gunman's firearms to him and told him the student union's peak hours. OpenAI says ChatGPT gave factual answers found across public sources and "is not responsible for this terrible crime."
Paris prosecutors have opened a criminal investigation into Elon Musk and X over alleged child sexual abuse material, plus non-consensual sexual deepfakes and Holocaust denial allegedly produced by Grok, and whether the deepfake controversy was deliberately amplified to boost the value of X and xAI. Musk and former chief executive Linda Yaccarino did not appear for voluntary interviews.
A US federal judge refused to dismiss writers' claims that Nvidia trained AI models including Megatron and Nemotron on The Pile, a dataset containing Books3, a collection of nearly 200,000 pirated books. Claims that Nvidia gave customers tools to download the data may also proceed; one alleging it fully controlled customers was rejected. Nvidia denies the allegations.
AI-generated avatars, mostly blonde women in swimsuits or military fatigues, are flooding TikTok, Instagram and Facebook with pro-Trump messaging ahead of November's midterms, AFP reports; who runs them is unclear. One Instagram soldier persona amassed nearly a million followers before being revealed as AI and suspended. A Purdue researcher says "financial profit" is likely the primary goal.
A Reddit user says strangers seeking lawyers, locksmiths and other services have called them for over a month, each saying Google's AI gave out the number, apparently as a placeholder contact. The user's removal request to Google has gone unanswered, they say; Google says it has safeguards and removal tools.
Canary Islands president Fernando Clavijo sent Health Minister Mónica García an AI-generated report saying rats can survive in water for up to three days, arguing against the hantavirus-hit MV Hondius staying at Granadilla port, according to elDiario. Health state secretary Javier Padilla dismissed the idea of an infected mouse swimming 200 metres to "colonise Tenerife".
The Delhi High Court granted Shark Tank India judge Aman Gupta an ex parte interim injunction barring defendants from using his name, image, voice or likeness, including via AI deepfakes, and ordered Google and other intermediaries to remove infringing URLs. His suit alleges fake endorsements, impersonation accounts, AI chatbot services and pornographic links.
An Observatório Lupa study of more than 289,000 posts on X from 24 March to 30 April 2026 found a disinformation campaign, which it says right-wing politicians coordinated, against Brazil's misogyny bill, PL 896/2023, including AI-generated fake videos claiming business owners had started firing women to avoid lawsuits under the future law.
