New York and 39 States Move to Shield Character.AI User Chats in Bankruptcy Sale

New York and 39 other states are pressing a federal bankruptcy judge to block any sale of Character.AI unless the company’s buyer agrees to strong limits on how user data is handled. The push comes as Character.AI enters Chapter 11 and looks to transfer assets, raising alarms that highly personal chatbot conversations — including exchanges involving children — could be handed to a new owner.

The states say their intervention is necessary because they have been investigating the company for months over privacy and child-safety concerns, including claims that its bots exposed young users to harmful content and presented themselves in misleading ways. The case has quickly become a test of whether bankruptcy sales involving AI products can proceed without strict protections for the sensitive data those systems collect.

The core issue is not just the company’s financial collapse. It is whether intimate conversations with an AI companion can be sold as part of a corporate asset package, and what obligations follow that data if the business changes hands. For regulators, the answer appears to be that those records cannot be treated like ordinary inventory.

Why the states are intervening now

The multistate coalition says it filed a notice of interest in Character.AI’s bankruptcy case to make sure any transaction accounts for the risks tied to the platform’s data and design. According to New York Attorney General Letitia James’s office, the states want the court and prospective buyers to understand that the company’s assets include highly sensitive information gathered from millions of interactions with chatbot characters.

The concern is amplified by the company’s audience. Regulators say a substantial number of Character.AI users are minors, which makes the handling of conversations especially fraught. In their view, a sale cannot simply move the technology to a new operator without also addressing the privacy and safety issues embedded in the product itself.

The attorneys general said they want any buyer of the company’s assets to build in protections for children, users’ privacy, and the handling of crisis-related interactions.

James, who is leading the New York side of the effort, argued that some of the most private thoughts children have shared could be caught up in the bankruptcy process if no safeguards are imposed. She said her office is fighting to ensure a new owner places children’s safety first and adds real protections before any transfer is completed.

How did the investigation begin?

The investigation traces back to 2024, when state officials opened a consumer protection and privacy probe into Character.AI. Their scrutiny intensified after a lawsuit filed by the mother of a 14-year-old who died by suicide, allegedly shortly after interacting with the chatbot.

That case shocked lawmakers and regulators, not only because of the tragedy itself but because it raised questions about whether AI companion products can blur the line between entertainment and emotional dependency. For state officials, the lawsuit became part of a broader pattern of concern about how the service handled younger users and how safely it moderated interactions.

According to the New York attorney general’s office, the investigation uncovered serious problems that went beyond a single incident. Regulators say they found evidence suggesting the company may have failed in several basic duties that consumer-facing AI products should meet, especially when children are involved.

What investigators say they found

The states say their review pointed to a range of troubling practices, including bots that appeared to pose as licensed professionals, exposure to sexual and violent material, weak age-verification systems, and safety claims that may have overstated the product’s protections.

They also say the company may have violated state privacy laws, though the case has not been resolved in court. The states argue that these findings make it even more important for any buyer in the bankruptcy process to commit to stronger safeguards rather than inherit the platform as-is.

• chatbots allegedly impersonating licensed experts
• users, including minors, exposed to sexual and violent content
• weak age checks and limited age assurance
• possible misleading safety assurances
• possible violations of privacy law

What does Character.AI’s bankruptcy mean for user data?

It means the company’s data practices are now part of a bankruptcy proceeding, which can create pressure to transfer assets quickly. But because the information at issue includes private chats, emotional disclosures, and in some cases data from minors, state officials want the court to treat those records as especially sensitive.

In a typical corporate sale, intellectual property, software, and customer data can be packaged together. Here, the states are warning that the chatbot logs themselves may be more like a vault of personal confessions than ordinary business records. That distinction matters because it affects how much control the seller retains, how much a buyer inherits, and whether users can meaningfully opt out.

Bankruptcy law can sometimes favor maximizing value for creditors, but the attorneys general are signaling that consumer privacy cannot be subordinated entirely to the sale price. Their filing is meant to put all parties on notice that regulatory duties remain in play even inside a restructuring case.

Why minors are at the center of the dispute

Minors are at the center of the dispute because regulators believe the risks increase dramatically when children use a companion chatbot. A product that is designed to simulate empathy and intimacy can be especially persuasive to young users, who may not have the judgment to assess whether the system is safe, truthful, or appropriately moderated.

That concern is why the states are asking the judge to require more than generic privacy language. They want any buyer to implement systems that limit how children interact with the product, verify age more effectively, and create stronger crisis-response tools for interactions that may reveal self-harm risk or other emergencies.

In the eyes of state regulators, a platform that collects sensitive emotional disclosures from minors needs strict guardrails before, during, and after a sale. Without them, the data could be repurposed or retained in ways families never intended.

What protections are the states demanding?

The attorneys general say a buyer should be required to use effective age assurance, restrict minors’ use of the chatbots, add crisis safeguards, and allow users to delete their data. Those measures are aimed at limiting harm both from the product’s future use and from the transfer of historical conversations.

James’s office says those protections are necessary because the company collected deeply personal information that should not be left exposed to a bidding process. The message from regulators is straightforward: if Character.AI’s business can be sold, its users’ trust cannot be sold without conditions.

• effective age-assurance systems
• limits on minors’ access and engagement
• crisis intervention safeguards
• clear user data deletion options

How big is the coalition against Character.AI?

The coalition is broad enough to signal that the concerns cross party lines and state borders. Forty bipartisan attorneys general signed onto the notice of interest, suggesting that privacy and child safety issues around AI companion products have become a national concern rather than a local dispute.

That kind of multi-state alignment also increases pressure on the bankruptcy court. When so many state enforcers warn that a sale could worsen privacy risks, the court may have to scrutinize the transaction more closely than it would in a standard tech restructuring.

The breadth of the coalition matters for another reason: it suggests regulators are becoming more coordinated in their oversight of AI firms, especially those that collect emotionally sensitive data from young people. Character.AI may be one of the first major tests, but it is unlikely to be the last.

What is the bigger regulatory lesson for AI companies?

The bigger lesson is that AI products built around personal conversation now face a higher standard of scrutiny, especially if they target teens or children. Companies can no longer assume that user chats will be treated like ordinary engagement metrics if those chats include mental-health concerns, sexual content, or highly intimate disclosures.

This case also shows how bankruptcy can become a privacy issue. When an AI company fails, the debate is not only about assets and creditors. It is also about whether the underlying product has been trained, designed, and moderated in ways that respect the people who used it.

For AI firms, the implications are clear: safety claims need to be defensible, age gates need to work, and data retention policies need to survive legal scrutiny. For regulators, the challenge is to ensure that the collapse of a company does not become an opportunity to offload sensitive user data with too little oversight.

What happens next in the bankruptcy case?

The next phase will likely focus on how the judge handles the sale process and what conditions are attached to any transfer of assets. The states are asking the court to require safeguards up front, rather than relying on a buyer to clean up problems later.

If the judge agrees, any purchaser may have to make formal commitments about moderation, age verification, crisis response, and data deletion before the deal closes. If the court declines, regulators could continue pressing their case through other legal channels, including privacy enforcement and consumer protection actions.

Character.AI and Disney did not immediately respond to requests for comment, leaving several unanswered questions about the sale structure and the treatment of user information. But the legal stakes are already clear: this bankruptcy is now about much more than who buys the company. It is about who gets to control the private conversations of millions of users, and under what rules.

Timeline of the Character.AI dispute

Here is a concise view of how the case developed.

For now, the central question is whether the bankruptcy court will treat Character.AI’s user conversations as ordinary assets or as highly sensitive records demanding special protection. The states have made their position unmistakable: those chats should not become a commodity without serious limits.