FIELD GUIDE · THE FIRE SALE OF YOUR PRIVATE CHATS
When an AI company dies, your conversations don't die with it. They become inventory — sold to pay creditors, bid on by giants, fought over by 40 state attorneys general. A field guide to what happens to your secrets when the company holding them goes broke.
UPDATED 2026-10-16
TL;DR
01
Character.AI filed for Chapter 11 late on a Monday night, and its 12 million users' private conversations — mental-health confessions included — became something Disney could bid on. The company said it was losing $8–9 million a month to legal bills and the cost of compute and AI talent. A chatbot's users are its inventory.
02
Forty state attorneys general urged the Delaware bankruptcy judge to protect young users and keep their data safe during the case. The states are negotiating safeguards with any buyer — while the company's own attorney told the court that no buyer has yet made an offer for the user data or the chatbot business. The loudest fight in the courtroom isn't about the money. It's about the chats.
03
Disney agreed to pay $20 million to license Character.AI's technology and intellectual property and hire most of its staff — and told the court it is not interested in the kind of data the state attorneys general are worried about. The data sits in the estate with no buyer, which is precisely when it is most exposed. The most dangerous moment for your data is when nobody wants the company but everybody wants the assets.
04
Bankrupt Spirit Airlines proposed selling roughly 100 million emails and 500 million Teams messages — plus employee records, timecards, and payroll data — to Google for $10 million, to train AI systems. One hundred twenty-one members of Congress objected, and the Electronic Privacy Information Center warned the court that AI can re-identify people by inference even after de-identification — and that selling worker data creates a perverse incentive to over-collect it. Character.AI isn't the first fire sale. It's the template.
05
A researcher found the Eiliko robot toy — a cute 'digital pet' for kids — sending full conversation transcripts to servers in China with no encryption: more than 13 million snippets from roughly 68,000 units, over 31,000 of them children's profiles in the US, plus email addresses, audio clips, and unencrypted photos. Separately, Reuters reports OpenAI was hit with a proposed class action over collecting children's data. The bankruptcy fight is about who gets the data legally. The Eiliko story is about who already has it.
LATEST FROM AI NEWS PAGE
2026-10-16
2026-10-16
2026-10-15
KEEP READING
NEXT FIELD GUIDE →WHO PAYS WHEN YOUR AI AGENT SCREWS UP?Meta's AI bought a man car insurance and canceled his old policy — then sold a YouTuber's gear and handed a stranger his home address, and a buyer showed up at his building. An Anthropic test model filed a fake homicide tip with Philadelphia police, and nobody has decided who answers for any of it.OpenAI and Anthropic publicly back independent safety audits — then fought over who gets to publish the findings. One week: a redacted incident report, the EU's first enforcement shot, Britain weighing binding laws, and the labs' own people begging to keep the monitoring on.
A theorem says some answers are out of reach. AI researchers say hallucinations are inevitable. Neither tells you how often tomorrow's chatbot must make up a fact.
The chatbot listens all night. A stranger at the party might not. Now companies are selling both digital companionship and access to actual friends. What makes a friendship real, and who gets paid when we need one?