Character.AI said it has filed for bankruptcy in the U.S. and agreed to sell some assets to Walt Disney for $20 million. It faces lawsuits from Kentucky, which says its chatbots encouraged users to cut, starve or kill themselves, and from Pennsylvania, which alleges they posed as doctors. Earlier this year it and Google settled a Florida mother's suit over her son's suicide.
An Anthropic report cited by the Daily Mail says Yemen's Houthi rebels ran three weapons programmes using Claude, including a multistage ballistic missile with an intended range above 1,500 miles. Claude reportedly helped with guidance software and fixes after misses. Anthropic says its safeguards blocked many but not all requests, and that it has evidence the group built offline simulation tools not relying on Claude.
Sharonville, Ohio, police Lt. Walter Cordes was indicted on a fifth-degree felony charge of unauthorized use of a computer, cable or telecommunication property. Prosecutors say he used the Flock camera system 29 times between Dec. 19, 2024, and April 20, 2026, to track a person he knew for personal reasons, and later admitted the searches. A Flock feature that flags suspicious searches surfaced the activity.
Superior Electoral Court judge Marco Aurélio Bellizze ordered Attorney General Jorge Messias to remove, within 24 hours, an AI-made Instagram video backing President Lula's reelection. The order partly grants a complaint from rival candidate Flávio Bolsonaro. The judge preliminarily found that the unlabelled synthetic characters breached disclosure rules, and left possible deepfake violations and a fine for the final ruling.
Pakistan's National Cyber Crime Investigation Agency in Islamabad registered a case against a man after an inquiry. According to the allegations, he expressed an intention to kill his former female friend and sought information on murder methods using AI and digital platforms. An agency official said the allegations remain under investigation and that the facts would be established through the legal process.
Gendarmerie units detained nine suspects in simultaneous raids across Düzce, Ankara, Adana, Antalya, Osmaniye and Manisa over alleged fraud using AI-made 'stock investment advisory' ads on social media that promised high returns, according to the Düzce governor's office. A court remanded all nine in custody. Investigators found that 805,000 lira had been transferred to their bank accounts.
A Joint Economic Committee inquiry launched by Sens. Maggie Hassan and Marsha Blackburn found that most of Tinder's machine-learning scam-detection models had not been updated in years, according to internal documents. It also found that nearly 1.5 million scammers or other bad actors on Tinder and Hinge matched with someone in the first nine months of 2025. Match Group has since deployed new models.
Actors Cláudia Raia and Jarbas Homem de Mello sued Facebook Brasil over AI-manipulated images of their 3-year-old son. The images added garments of African-derived religions to suggest a candomblé initiation, and the couple say they spread on Instagram and Facebook and drew public attacks. They seek removal of the images, the posters' registration data and R$40,000 in moral damages.
On Oct. 7 the Rio de Janeiro bar association (OAB-RJ) obtained a federal appeals decision barring Resolve Juizado from charging about R$20 for AI-generated initial petitions produced without a lawyer's involvement. The bar argues the service misleads consumers and may amount to illegal practice of law. It says the ruling targets the paid sales, not the use of AI in legal drafting.
Judge Lauren King of the Western District of Washington sanctioned plaintiff Jtwain King for citing nonexistent cases in a brief opposing Equifax, rejecting as not credible her claim of mere 'citation verification errors'. She must state under penalty of perjury whether generative AI was used in all future filings and correspondence with opposing counsel. The court declined to dismiss the case or award fees.
A justice of the peace in Russia's Amur region fined a Magdagachi woman 15,000 rubles ($179) for a ChatGPT-generated photo of a bear roaming a local street, which she made as a joke for a friend. The image went viral, neighbours called police and authorities summoned hunters. The court reduced the 30,000-ruble statutory fine because of her confession and remorse.
A 29-year-old from Naples who had repeatedly failed the driving licence exam was caught by an examiner at Avellino's Motorizzazione Civile. The candidate was using a mobile phone with an AI system that read the quiz questions and supplied answers. State Police seized the device and reported the candidate. Investigators are seeking whoever supplied the device for an agreed €1,000.
