Character.AI Says It Has Filed for Chapter 11 Protection

Follow the money in this bankruptcy and you learn who matters in the AI economy. Walt Disney gets a perpetual license to Character.AI's intellectual property for $20 million. The lonely teenager who confided in a chatbot gets a notice that her transcripts may be disclosed to prospective buyers.

Character.AI says it is in Chapter 11 because of "legal, regulatory and operational challenges." That is a polite way to describe a product accused of harming children. The company faces lawsuits from families, including the mother of a 14-year-old Florida boy who died by suicide after months of intense conversations with one of its bots. These are allegations, and courts will weigh them. But the business model was never mysterious. Build an emotionally persuasive companion, make it sticky, let minors in, and treat the consequences as someone else's problem. Only after the lawsuits, the public outrage and federal scrutiny did the company restrict open-ended chats for under-18 users.

Now comes the familiar ending. The founders and much of the technical talent were absorbed by Google in a 2024 deal reported at around $2.7 billion. What remains is a shell with a user base, a pile of liabilities and a court-supervised sales process. Bankruptcy is a legitimate tool. It is also a very good way to turn a moral reckoning into a claims queue, where a grieving family becomes one creditor among many and the most valuable asset is whatever can be sold fastest.

That asset, apparently, includes us. The company says chat transcripts, audio recordings, images and videos could be disclosed during sale negotiations. Think about what people tell a companion bot: their fears, their sexuality, their suicidal thoughts, their secrets from parents and spouses. We have a name for that kind of material when a human therapist holds it. Here it is a line item in a data room, protected by a privacy policy the company wrote for its own convenience and can lawfully point to as consent. Nobody reads those policies, and "consent" is a legal fiction when a lonely 15-year-old clicks "agree."

We have seen this before. When 23andMe went bankrupt, millions of people suddenly had to worry about who would own their genetic information, and state attorneys general scrambled to object. Congress did nothing durable. The lesson was obvious: in America, the most intimate data we produce is treated as ordinary property, sold off when the owner runs out of runway.

The Disney detail should embarrass everyone who says AI regulation is impossible. Corporate lawyers can negotiate a perpetual license for fictional characters in a matter of weeks. The system can defend a cartoon mouse with precision, yet it still has no clear rule that a child's confessions to a machine cannot be auctioned. The law is perfectly capable of protecting property. It has simply chosen not to protect people with the same energy.

Democrats should not hide from their share of this. Too many in our party spent years treating Silicon Valley as an economic partner and a donor base, then acted surprised when the "move fast" culture ran over children. Hearings and sternly worded letters are not policy. We need a federal baseline privacy law with real teeth, a ban on selling or transferring sensitive conversational data in bankruptcy without affirmative opt-in consent, strict age-assurance and design duties for companion bots, and a court-appointed independent privacy advocate in cases like this one. The FTC and state attorneys general should be in that courtroom now, not after the data has changed hands.

And the families should be at the front of the line, not the back. If a company's product allegedly contributed to a child's death, the sale should not close without a serious accounting of who pays for that harm. Disney will get its license. The question is whether everyone else gets anything at all.