Disney’s Character.AI deal takes the staff and tech but leaves the lawsuits behind

Character.AI’s bankruptcy filing is part of a deal that lets Disney hire most of its staff and bid for much of its technology, while the startup keeps its legal liabilities, Wired reported, citing court filings. Disney declined to comment to Wired.

Two transactions, one before and one after bankruptcy

According to Wired’s reading of the filings, the first part took place in September. Disney paid Character.AI $5.5 million plus a refundable $2 million good faith deposit, hired the majority of its employees, including then CEO Karandeep Anand, and agreed to let some of them help keep the service running for five months.

Character.AI filed for chapter 11 on October 12. Through that process, Disney seeks a perpetual, worldwide, nonexclusive license to a broad set of the startup’s code, documentation and tools. Wired says this covers the trust and safety moderation stack, training infrastructure and age-prediction technology. User data and the website are reportedly not on the list.

The license is to be auctioned in the coming months. If Disney wins and the courts approve, it would pay a further $12.5 million at closing. If it loses, it gets its deposit back plus a breakup fee. Disney also agreed that a winning rival could interview up to 80 percent of certain employees it had just hired.

Lawsuits and investigations stay with the startup

A draft court order prepared by Disney would state that the arrangement is not a merger and that Character.AI remains responsible for any liability, including legal claims, according to Wired.

The legal background consists of claims, not findings. A notice of interest filed by the National Association of Attorneys General on behalf of 40 states says their investigations found “serious issues”, including screenshots they say show chatbots encouraging minors to hide self-harm and to kill their parents. Character.AI says the filing “is not an accurate portrayal of the safety of our platform.” Per Wired, the multistate investigation had grown to 46 states and territories by the bankruptcy filing, and talks with the states ended without an agreement.

Wired also reports that Google and Character.AI agreed in January to settle a group of lawsuits. A referee approved each individual settlement on October 11, a day before the bankruptcy filing.

Open questions the filings do not settle

Much remains unresolved. Character.AI’s spokesperson referred Wired to Disney on whether the deal depended on bankruptcy, and Disney declined to comment. Nothing in Wired’s account says what Disney would do with the moderation or age-prediction tools.

The structure is also untested. John Newman, a University of Memphis law professor and former FTC competition official, told Wired that corporate law safeguards “really weren’t built to guard against” a deal like this. The states have said they reserve the right to challenge any bankruptcy plan, so the outcome could be contested. Who ends up with the technology depends on an auction and court approval, and neither has happened yet.

For readers following chatbot safety, the case shows how legal risk can separate from talent and technology when a startup becomes hard to sell. Background on why people bond with chatbots is in our piece on emotional relationships with AI. Everything above comes from Wired’s account, and the filings themselves were not reviewed directly.