Spirit Airlines Data Sale: No Seat for the Authors | Elacity
Spirit Airlines' bankruptcy estate is selling about 100 million work emails and 500 million Teams messages as AI training data. The people who wrote them are not a party.
The Spirit Airlines Data Sale Has Bidders, a Judge and No Seat for the Authors
The emails and chat messages you write at work can outlive your job, your employer and any say you had over them. The Spirit Airlines data sale, due back before a bankruptcy judge on November 6, is the clearest proof yet: years of workplace conversation are being sold as AI training data, and the people who wrote it are not a party to the deal.
At an August auction, Spirit's bankruptcy estate selected Google as the winning bidder at $10 million, with Mercor as the alternate. The asset is what the filings call deidentified data. Rep. Steven Horsford's office says court findings put it at roughly 100 million emails and 500 million Microsoft Teams messages, plus employee records, payroll and tax information, created by workers as a condition of their jobs.
Then an AI startup, micro1, offered $12.5 million for the same records of a carrier that went out of business this year. When AI companies bid against each other for an airline's office chatter, the value of ordinary human work has moved somewhere its authors cannot see.
The Failure of the Bankruptcy Data Brake
The Bankruptcy Code does have a brake for personal data. Section 363(b)(1) applies when a company told the individuals it sold to that it would not transfer their personal information. A sale that breaks that promise needs a court hearing with a consumer privacy ombudsman first.
That brake was built for customers. In Spirit's case, the asset schedule marked 97.5 million passenger profiles as not included, while the workers' own words stayed in, with deidentification as their protection. Google later agreed to drop some datasets, including timecards, the same report notes.
The flight attendants' union objected that stripping names does not make confidential employee records safe. More than 100 members of Congress warned both companies that removing direct identifiers "does not necessarily make a dataset anonymous," and Google told The Record it is not looking to buy any personal information.
Look at what each protection is: an objection, a letter, a deidentification agent, a judge's discretion. Every one asks the current holder to behave. None gives the authors anything to hold.
A fair reading concedes a lot. Work email has always belonged to the employer, nothing here suggests the estate is acting outside the law, and the court may still narrow the deal or attach conditions. The problem is structural: whoever holds the readable copy holds the asset, and an estate sells whatever it can open.
The Elacity Solution: Make the Owner Part of the Asset
Elacity starts from the opposite default. AI is turning human output, the conversations, documents and judgment machines learn from, into an input worth bidding for. If that is true, the person who created the input should hold it as property, not appear in a sale only as a privacy risk.
Sasha Mitchell, Elacity's founder, puts it plainly: "The people who create the value should own it. That is the entire reason Elacity exists." Here is what we are building so the lever sits with the owner, and where it is not switched on yet.
1. The data carries its owner's terms
Through Elacity dDRM, a dataset, document or body of work is designed to become a Wealth Capsule: an encrypted good with its terms recorded alongside it, covering who may use it and at what price. A buyer would acquire rights under those terms, not a raw copy to use however it likes.
2. The key answers to rights, not possession
Holding the file is not meant to be enough. The key that opens a Wealth Capsule is designed to be split across a 2-of-3 quorum of separate machines, each re-checking the holder's on-chain rights before releasing its share. That path is built but not switched on yet, so read it as the design, not a guarantee.
Had those emails and messages been sealed from the start, a new holder would get content it cannot open unless the recorded rights let it in. Who holds those rights, and whether authors are among them, is a choice made when the content is sealed.
3. Your machine is the source of truth
Beneath it sits ElastOS, the open runtime Elacity helps build, where your own machine is the source of truth and the cloud is a guest. Its design keeps data in the owner's space: code comes to the data, and crossings between spaces are meant to need the owner's signed permission.
Part of that is real today. In ElastOS, an app can't make its own permission: it asks, and the Runtime mints and signs a capability token, then checks it before the app can act. Capabilities, not access control lists: instead of the owner keeping a list of everyone allowed in, a holder presents a narrow, signed credential that is checked before it is used, and by design authority starts at zero and is granted narrowly.
What this does not fix
None of this rewrites employment law. Records a company creates on its own systems will keep belonging to the company. The lesson is for everything you create on your own terms, which mostly sits in someone else's database today:
- the work you publish or sell, from writing to code to music;
- the datasets you build, which AI developers increasingly pay for;
- your health and personal records, the subject of our look at the price of a medical record.
For the full mechanism, read how decentralized DRM decides who holds the key. More on who holds power over data lives under Social Architecture.
Ask who could have said no
When the Spirit sale returns to court, the argument will be about what the buyer may do with the data. The better question is who could have refused. Follow Elacity on X for the next piece on who holds the key to the work you create.