A federal judge in California on 25 September preliminarily approved Anthropic's $1.5 billion settlement with authors who argued it unlawfully used millions of pirated books to train its assistant Claude. Judge William Alsup, who ruled in June that saving more than 7 million pirated books to a "central library" violated their rights, will decide on final approval after authors can file claims.
Police in Incheon said on 25 September they are investigating a male high school student accused of using AI deepfake technology to put the faces of female high school students onto images with sadistic sexual implements and spreading them on social media. Four victims filed complaints; one victim's father said his daughter hyperventilated and was taken to hospital by ambulance.
Alameda County Superior Court Judge Victoria Kolakowski dismissed an $8.7 million tenants' dispute lawsuit after finding the plaintiffs submitted a deepfake video and at least two other pieces of evidence altered by generative AI. The sanction, the judge wrote, serves as a deterrent "showing the public that the Court has zero tolerance with attempting to pass deepfakes as evidence".
Kateryna Izotkina, a 24-year-old Ukrainian, was remanded in custody in Malta after pleading not guilty on Wednesday to fraud, money laundering, illicit gains and organised crime charges over a cryptocurrency scam promoted with an AI-generated video of Prime Minister Robert Abela. One complainant says she lost €52,000; police arrested the accused at a controlled delivery using counterfeit notes.
Who did it: HYBRID AI Hallucination Cases database
A federal judge in New Jersey on 25 September granted summary judgment against a plaintiff suing police over his arrest, then ordered his two lawyers to show cause after finding inaccurate quotations and citations on nearly half the brief's pages, including Westlaw citations that do not exist. The judge said this raised concern the brief was possibly prepared using generative AI without proper verification.
Who did it: HYBRID AI Hallucination Cases database
In a 25 September ruling, the Frankfurt Regional Court said three verbatim quotations from Federal Court of Justice decisions in a plaintiff's lawyer's brief were, as the lawyer had to concede, complete fabrications with nonexistent references, dates and file numbers. The court hoped a chatbot, rather than the lawyer, had "hallucinated" them, noting its own chatbot queries returned the correct legal view.
French TV host Karine Le Marchand warned on Instagram on 25 September that videos circulating on WhatsApp, apparently AI-made, use a smoothed version of the host's face and an altered voice to ask people for money, including €3,500 supposedly for a charity. "I would never ask people I don't know for money," the presenter said.
MGM China Holdings warned on Thursday that an AI-generated deepfake video falsely depicts its chairperson and executive director, Pansy Ho, promoting a fake investment plan with promises of massive returns. The company said neither it nor Ho had any involvement, vowed to pursue legal action against those responsible and urged the public not to share the video.
Elon Musk's xAI has sued OpenAI in California federal court. The complaint alleges a 'deeply troubling pattern' of hiring away former xAI employees, including two engineers and a senior finance executive, to obtain trade secrets related to Grok, its source code and its data center operations. OpenAI called the suit 'the latest chapter in Mr Musk's ongoing harassment.'
