South Korean authorities have arrested Kim Se-ui, who runs the YouTube channel Hover Lab, after the Seoul Central District Court granted a warrant. Police say a recording he posted, in which a late actress appears to say she had dated actor Kim Soo-hyun since middle school, was AI-generated, and allege he manipulated text-message screenshots. He denies the charges.
Berlingske has sent home an employee after an internal AI tool, used against the paper's AI guidelines, invented three professors and their quotes in an article on economic advisers favouring a pesticide ban over drinking-water treatment; the piece also misquoted the advisers twice. Editor-in-chief Tom Jensen said the case weakens Berlingske's credibility; an external consultant will review the employee's articles.
Stephen Buckley, 47, was convicted and fined €400 at the district court in Tralee after pleading guilty to possessing four AI-generated child pornographic images, including one manipulated with an AI app to undress a teenage girl, and three cartoon-like videos. He is believed to be the first person in Ireland convicted of possessing AI-generated child pornographic material.
Spain's Guardia Civil is investigating a man over alleged sexual offences after digitally manipulated images placed recognisable minors' faces on naked bodies, prompting four complaints in a town in Cáceres province. The Guardia Civil says he made them with an AI app, using photos taken mainly from the Instagram profiles of victims he was friends or contacts with.
The New York Times Tech Guild has filed grievances alleging management breached its contract by using AI tools DX and Glean to track and evaluate staff; the union says DX statistics have been cited in disciplinary conversations. It and the Times Guild filed unfair labor practice charges alleging the paper refused to answer requests about its AI use. The Times disputes the characterisations.
Authors have filed a proposed class action in New York federal court against Meta Platforms, CEO Mark Zuckerberg and, unusually, two former senior AI researchers named individually; it alleges the researchers carried out the company's mass piracy of tens of millions of books, in the Llama copyright case.
Who did it: HYBRID AI Hallucination Cases database
The Oklahoma Supreme Court publicly reprimanded lawyer Matthew Brett Reeves, who admitted adding five unverified ChatGPT citations, which a federal court called “completely made up”, to two motions filed for the Alabama Department of Corrections. The reciprocal discipline follows a federal sanctions order and reprimands or censure in Texas, Tennessee and Alabama; the court found he had knowingly made false statements of law.
Presidential media aide Bayo Onanuga said social media critic Martins Vincent Otse, known as Verydarkman, should face the full weight of the law for spreading an AI-generated audio of President Bola Tinubu, after an X user posted that it was played in his video. A Daily Post check found no such audio in his videos.
Mobile game Tower of Saviors apologised after online users noticed a shot in the promotional video for its new character Black Gold Genghis Khan was almost identical to an animated attack from Wuthering Waves. Its makers admitted using AI-assisted production, pulled the video, began an accountability process and pledged to ban using third-party works as a direct basis for promotional material.
