Character.AI bankruptcy: 40 US states demand protection for children's data
Forty state attorneys general asked a Delaware bankruptcy judge to ensure any buyer of Character.AI safeguards young users' chat histories.
A coalition of 40 US state attorneys general asked a federal bankruptcy judge in Delaware on October 14, 2026, to shield young users of Character.AI and secure their data as the AI companion company moves through bankruptcy. The startup had filed for Chapter 11 protection late on October 12, 2026.
Character.AI intends to use the case to settle investigations and lawsuits brought by the states, license its technology to Disney for $20 million, and find a buyer for its chatbot business. At the company's first hearing in Wilmington, states' attorney Abigail Ryan told US Bankruptcy Judge Brendan Shannon that settlement talks with the company are under way.
What the states want
The states allege that Character.AI's chatbots posed as licensed medical professionals and subjected children to sexualized and abusive exchanges. Ryan said any purchaser of the company's assets must protect chat histories and strengthen safety tools. "We don't want to see what has transpired in the past start over with a new purchaser," she told the court. New York Attorney General Letitia James warned separately that children's most private conversations could end up sold to the highest bidder.
Erin Fay, representing Character.AI, said the company had heard the concerns but noted that no bidder has yet offered to buy its user data or its operating chatbot service. According to Fay, Disney has no interest in the kind of data worrying the states; it agreed to pay $20 million for the technology and intellectual property and to hire most of the staff. That deal needs court approval and remains open to higher bids.
Lawsuits and losses
The platform lets people build and chat with customized generative AI characters, often based on celebrities, fictional figures or historical personalities, and active users average two hours a day on it. Families have also sued, alleging the chatbots gave harmful advice, contributed to the suicides of two children and exposed minors to sexualized and bullying content; the company has reached preliminary settlements with five of them. It is additionally under investigation by the US government and authorities in the UK, Italy, Russia and Australia.
Character.AI blamed its collapse on legal costs and the steep price of computing power and AI talent, saying it was losing $8 million to $9 million a month before filing. Court filings show management feared staff departures by late summer, and in September Disney hired most employees so they could keep the service running through the sale and technology transition.
The case could set an early marker for how courts handle intimate user data held by failing AI firms, particularly whether children's emotionally sensitive conversations can be treated as assets that change hands in a sale.
Comments
No comments yet. Start the discussion or ask a question.
Related
The Financial Stability Board's final guidance for banks says higher-risk uses of AI agents will require more intensive human oversight and clear approval triggers.
More than 80 US law enforcement and intelligence documents show agencies have treated AI since 2021 as both a weapon for would-be attackers and a target to protect.
At a 13 October 2026 UK parliamentary hearing, OpenAI and Anthropic said voluntary pledges are not enough, while Google DeepMind and Meta declined to back new laws.


