An unauthorized AI-generated video showing Scarlett Johansson and over a dozen other Jewish celebrities, including Drake, Jerry Seinfeld and Steven Spielberg, opposing Kanye West has gone viral. Johansson denounced "the misuse of A.I., no matter what its messaging" and urged the U.S. government to make passing legislation limiting A.I. use a top priority.
Louisiana State Police say a routine phone inspection of Ryan Glaser, 40, a convicted sex offender, turned up child sexual abuse material, and he was charged with four counts of possessing unlawful deepfakes in July 2024. Investigators say Gregory Duncan Jr., 24, created and distributed the material to him; Duncan was charged with distribution and booked on Feb. 4, 2025.
Portugal's Polícia Judiciária detained a 22-year-old it says was responsible for developing WormGPT, a ChatGPT-like model "without the limitations of the original" allegedly sold on hacking forums for illicit activity. Police say they seized the system's entire computing infrastructure during operation "Neutral Kill Switch"; the suspect is to face a first judicial interrogation.
Brazil's Attorney General's Office (AGU) sent Meta an extrajudicial notification asking it to remove within 24 hours an AI-manipulated video showing a fake interview with Vice President Geraldo Alckmin about the use of public money. The AGU warned the platform could be held liable if it does not comply, and asked that the video otherwise be labelled as AI-generated.
Asked whether Uyghurs face genocide, DeepSeek's chatbot reportedly dismissed the claim as "severe slander" and "completely unfounded," repeating China's official stance that human rights concerns in Xinjiang are foreign interference. Uyghur activist Rahima Mahmut told The Independent that "the Chinese government is trying to erase the Uyghur people by employing AI to mislead the public."
Malta's Office of the President warned that several pages and profiles on Facebook, X and Instagram are impersonating President Myriam Spiteri Debono, using deepfake videos and fake photos to lure people into fake financial investments, mainly crypto. The office disassociated itself from the content and said it is immediately reporting the fraudulent pages to the relevant authorities.
Enroly, which assesses student interviews for universities, says that of approximately 20,000 interviews this January intake, 1.3% involved outright deception, including lip-syncing (0.6%), impersonation (0.5%) and deepfake attempts (0.15%), described as "fake faces layered over real ones." The company says deepfakes were the biggest surprise and that it has "already stopped several attempts."
A complaint filed with France's CNIL by lawyer Jérémy Roche alleges Mistral AI breaches the GDPR by using free Le Chat users' prompts to train its models without a clear way to object, short of paying for a subscription or emailing the company. Mistral told Le Figaro it has always allowed users to refuse such use.
