This week, Character.AI declared bankruptcy. While some saw the fall of a chatbot platform as a harbinger for the industry, court filings indicate that the decision was actually a key part of a deal that has let Disney hire most of Character.AI’s staff—including its now former CEO Karandeep Anand—potentially acquire most of Character.AI’s technology, and avoid the beleaguered startup’s many legal liabilities and obligations.

In the past few years, Character.AI has faced multiple private lawsuits over the safety of its chatbots, as well as investigations from regulators around the world. In a declaration filed as part of the bankruptcy, Character.AI’s newly appointed CEO explained that the goal of the bankruptcy was to reach a “consensual resolution” with personal injury claimants and a group of states investigating the company, while maximizing the value of the company’s assets for shareholders.

The Disney deal seems to be structured so that whatever is left of Character.AI will retain responsibility for those legal battles, some of which are still ongoing.

The move appears to be a new twist on what some have dubbed a “reverse acquihire,” a process in which a buyer carefully orchestrates a succession of deals with a target company, usually in an effort to hire key employees and obtain expansive license to use and build on the target company’s unique technology. Stripped of its employees, the target becomes a husk of itself.

Companies often take the position that reverse acquihires are not actual mergers, and that they do not have to notify federal antitrust regulators about those deals as they would otherwise be required to by law. These arrangements have become more commonplace in the AI industry in recent years, drawing the ire of Congress.

In the case of Character.AI, a motion filed by Character.AI’s legal counsel says that Disney required that the deal be “fully consummated” through bankruptcy proceedings because of the startup’s “diminishing cash reserve and significant litigation and regulatory risks.”A spokesperson for Character.AI referred WIRED to Disney for comment on whether the deal was contingent on the startup declaring bankruptcy. Disney declined to comment on this story.

According to court filings, the deal involves “two interrelated and integrated” transactions—one before Character.AI declared bankruptcy, and one after.

As part of the first transaction, which occurred in September, Disney paid the startup $5.5 million, along with a refundable “good faith deposit” of $2 million. Disney also hired the “majority” of Character.AI’s employees, including the CEO, and agreed that it would let some of the now-Disney employees help keep Character.AI operational for five months.

Character.AI filed a petition for chapter 11 bankruptcy on October 12. Through the bankruptcy proceedings, court filings show that Disney hopes to secure a perpetual, worldwide, nonexclusive license for a huge range of Character.AI’s code, documentation, frameworks, tools, and “other artifacts,” including those produced for its trust and safety moderation stack, training infrastructure, and age-prediction technology. The list of IP included in the filings does not include users’ data (which can get additional protections in bankruptcy proceedings) or Character.AI’s website.

That license will be auctioned off in the coming months. If Disney is chosen as the winning bidder by Character.AI and a committee of unsecured creditors—and courts approve the deal—it has agreed to pay Character.AI an additional $12.5 million as a closing payment. If Disney is not selected, the company will be refunded the $2 million deposit and receive a breakup fee.

In order to “preserve a competitive process,” Disney also agreed to permit the winning bidder to interview and potentially extend offers to up to 80 percent of “certain newly hired employees” that Disney already hired from Character.AI. Disney has the right to select which employees.

However, experts worry that bouncing Character.AI employees from one company to another in the span of months is unrealistic, and that the deal Disney and Character.AI have struck cannot be easily unwound.

In addition, Disney prepared a draft of a court order referred to as the “363 Free and Clear Order” in the filings, a version of which would be approved by the court at or before closing, that would clarify that this process should not be viewed as a merger of the two companies and that Character.AI would remain responsible for “any liability or other obligation” it might have, including any legal claims against it.

A spokesperson for Character.AI said they were unable to speculate on the outcome of the bankruptcy proceedings and referred WIRED to the docket.

“There's decades of corporate law that developed to deal with the problem of a buyer who always wants to take on all the assets and never wants to take on any of the liability,” says John Newman, a law professor at the University of Memphis School of Law and a former deputy director of the Federal Trade Commission’s Bureau of Competition. But the Disney-Character.AI deal is extremely unusual and structured in a way, he says, that “all those corporate law safeguards really weren't built to guard against.”Not Its First Rodeo This isn’t even Character.AI’s first reverse acquihire.

In August 2024, Google reportedly paid the startup $2.7 billion to license its technology. A spokesperson tells WIRED that that license was for the existing AI technology at the time of the deal, and that Google also hired Character.AI’s founders and other members of its ML pre-training research team, but that the “vast majority” of the startup’s employees remained at the company.

In the following years, Google was named alongside Character.AI as a defendant in multiple lawsuits, including one that alleged a 14-year-old boy committed suicide after Character.AI’s chatbot “cultivated an emotionally dependent and sexually exploitative relationship.”Google argued that because it was still a separate company from Character.AI, it was not responsible for the teen’s death, which it said occurred prior to Google’s licensing deal. The judge was not persuaded, determining that the plaintiff sufficiently alleged that Google was liable for aiding and abetting. Google declined to comment for this story.

In January of this year, Google and Character.AI agreed in principle to settle that and a slew of similar lawsuits. “Character.AI has taken innovative and decisive steps with regard to AI safety and teens, and will continue to champion these efforts and push others across the industry to adopt similar safety standards,” Character.AI said in a statement at the time.

According to an order of dismissal in one of the cases, the overall settlement was approved by a California probate court in July, and the court tasked a referee to hear out all the issues pertaining to each individual settlement. On October 11, a day before Character.AI declared bankruptcy, the referee approved each settlement.

The Tech Justice Law Project and Social Media Victims Law Center, which represented plaintiffs in many of the lawsuits, declined to comment.

Despite losing key staff after the Google transaction and, according to the current CEO, struggled with profitability “at all times,” Character.AI continued to develop AI tools, including “real-time audio-driven FaceTime-style video generation with fictional characters,” according to the current CEO’s declaration. In August 2025, Character.AI worked with an investment banker to try to find either a strategic buyer or investor.

According to the bankruptcy filings, Character.AI and the banker received “consistent feedback” that its product and user base were “impressive,” but that the surrounding litigation and “regulatory overhang” made it unbuyable. That November, Character.AI learned that a coalition of state attorneys general had started investigating the company.

When the company filed for bankruptcy, the number of states and territories involved in the multistate investigation had climbed to 46. According to a notice of interest filed in the bankruptcy proceedings by the National Association of Attorneys General on behalf of 40 states, those states’ investigations uncovered “serious issues” with Character.AI, including screenshots of Character.AI chatbots encouraging minors to hide their self-harm injuries and kill their parents.

A spokesperson for Character.AI says that the filing “is not an accurate portrayal of the safety of our platform.”

The filings say that Character.AI was in the process of negotiating with the states over how to address the issues they identified, but did not reach any agreements before declaring bankruptcy. If the deal is completed and approved by the courts, Disney will have both Character.AI’s staff and much of its technology stack, but will be able to sidestep the ongoing legal issues that contributed to Character.AI’s downfall, including the multistate investigation.

The states who filed a notice of interest in the bankruptcy have made it clear that they reserve their rights to challenge any bankruptcy plan, potentially setting up a fight over the ultimate outcome of the bankruptcy.

“It’s bad for society to let companies exercise control without responsibility,” says Newman.

“And that’s what it looks like.”