There is a long history of personal data becoming a contested asset when companies go bankrupt. But the Spirit Airlines Google data sale raises a new and much bigger question: what happens when decades of employee data and internal communications become an asset sought by an artificial intelligence (AI) company?
That question is now before a US bankruptcy court.
In August 2026, Google won a $10 million bankruptcy auction for a large collection of Spirit Airlines’ internal business data. The proposed transaction includes employee records, emails, Microsoft Teams messages, calendars, documents, spreadsheets, software code, operational records and other business information.
The data stretches back decades, with employee records dating to 1986. Court records reportedly identify more than 175,000 employee records, roughly 100 million emails and about 500 million Microsoft Teams records among the data covered by the transaction.
But one distinction is important from the outset: Google has not yet received the data.
The Spirit Airlines Google data sale still requires bankruptcy court approval. The hearing originally scheduled for August was postponed to September 9 after the Association of Flight Attendants-CWA, the union representing Spirit flight attendants, objected to the transaction. The union is seeking additional protections for employee information.
That distinction matters because the most important issue in this case is not whether Google has already obtained someone’s private records. It has not. The important question is whether a company’s historical employee data can become a commercial asset in bankruptcy and, if so, what privacy protections should follow that data after the company’s collapse.
The question is not entirely new.
Bankruptcy Has Already Tested Idea That Personal Data is Corporate Asset
One of the best-known examples dates back to 2000, when online retailer Toysmart.com filed for bankruptcy.
Toysmart had collected detailed information about its customers and had promised in its privacy policy that the information would not be shared with third parties. When the company attempted to sell its assets, it also sought to sell its customer database.
The Federal Trade Commission (FTC) challenged the proposed sale, arguing that the transfer conflicted with Toysmart’s privacy promises. The eventual settlement allowed a transfer only under strict conditions. The buyer had to be a qualified company in a related market, the data could not be sold as a standalone asset, and the buyer had to honor Toysmart’s privacy commitments. A future material change in the way the information was used required affirmative consent from customers.
Although these settlement conditions were established, no buyer met those criteria. Ultimately, majority owner Disney paid creditors $50,000 to destroy the database completely.
The case established an important principle: bankruptcy does not automatically erase the privacy commitments a company made when it collected personal information.
The issue appeared again in the 2011 bankruptcy of Borders.
Borders had collected information from more than 20 million customers, including purchase histories and email addresses. The FTC told the bankruptcy court’s consumer privacy ombudsman that the company had made privacy promises that restricted how customer information could be transferred.
The FTC recommended either obtaining customer consent or imposing substantial restrictions on any transfer. It specifically pointed back to Toysmart as a model, including restrictions on the buyer, limits on how the information could be used, and continued adherence to the original privacy commitments.
Ultimately, Barnes & Noble purchased Borders’ intellectual-property assets, including its customer list, for $13.9 million. The bankruptcy court approved the transfer with privacy protections, including an opportunity for Borders customers to opt out before their information was transferred to Barnes & Noble.
RadioShack produced another major example in 2015.
When RadioShack entered bankruptcy, its customer information was among the assets being offered for sale. The FTC said the database included names, addresses, email addresses and purchase histories and that the company had made privacy promises concerning that information.
Thirty-eight states subsequently participated in opposition to aspects of the proposed transfer. The eventual settlement sharply restricted what information could be transferred. Most of the consumer data was to be destroyed, while only limited information could be transferred under specific conditions and with restrictions on future use.
These cases demonstrate that the basic problem is not new.
Corporate bankruptcy can turn information that people provided in one context into an asset that creditors, buyers and courts must decide how to treat.
What makes Spirit Airlines Google data sale different is the nature of the information, the intended buyer and the role of artificial intelligence.
Spirit is Not Simply a Customer Database
The Spirit transaction is fundamentally different from the classic bankruptcy privacy disputes because much of the attention is focused on employee-generated and employee-related information.
The proposed Spirit Airlines Google data sale includes employee records, emails, Microsoft Teams messages, calendars, files and other internal business information. The court records also identify large quantities of operational and technical information, including software code and development data.
Spirit and Google have said the data will be “de-identified” before Google receives it. Google has said it will not receive personal information from the dataset. Passenger profiles and Free Spirit loyalty-program information are excluded from the transaction.
Those protections are important.
But they do not resolve the central issue raised by the flight attendants’ union.
The Association of Flight Attendants-CWA argues that de-identification and confidentiality are not the same thing. In its objection, the union points specifically to the requirement that the de-identification process preserve “referential integrity across the data set.” In practical terms, the datasets must retain relationships between records rather than simply becoming disconnected pieces of information. The union argues that preserving those relationships raises concerns about whether information concerning identifiable individuals or small groups could potentially be reconstructed.
That is an argument being made by the union, not a judicial finding that individuals can in fact be re-identified.
That distinction is important.
The bankruptcy court has not yet ruled in the Spirit Airlines Google data sale whether the proposed protections are sufficient.
De-identification and Confidentiality Are Different Questions
The Spirit dispute also highlights an important distinction that is easy to miss.
Suppose an employee’s name, employee number and other direct identifiers are removed from a record.
That answers one question: Can the record be directly linked to a named individual?
But it does not necessarily answer another question: Does the content of the record remain confidential?
The flight attendants’ union is specifically raising that second question.
Its objection says the proposed de-identification addresses whether records can be traced to named individuals but does not, by itself, determine whether the contents of those records remain confidential. The union also points to the requirement to preserve relationships among records across the dataset.
This is significant because workplace data can contain information that is sensitive even without a person’s name.
An employee’s communications, employment records, payroll information, scheduling information or workplace documents can reveal circumstances about that employee’s professional and personal life. The existence of such information within an employer’s systems is a fact of modern digital employment. The legal question is what happens to that information when the employer ceases to exist and the data becomes part of a bankruptcy transaction.
That question is much less settled than the earlier debate over customer databases.
The AI Dimension Changes the Stakes
Google has said the Spirit data could help improve its products and AI models.
That matters because the potential value of the dataset is not simply the information contained in individual documents.
The value can also come from the relationships among those documents.
- An email is one piece of information.
- A Teams conversation is another.
- An employee record is another.
- A calendar entry is another.
- A document, spreadsheet, workflow or operational record is another.
When those records remain connected, they can provide a much richer picture of how an organization operates.
That is one reason the preservation of relationships among records has become a central issue in the Spirit dispute. The same structure that can make a large enterprise dataset useful can also make privacy protections more complicated.
Again, it is important not to go further than the evidence allows. There has been no judicial finding that Google will use the data to identify individual Spirit employees, nor has the court determined that the proposed de-identification process is inadequate.
The issue in Spirit Airlines Google data sale is that the transaction creates a legal and technological question that earlier bankruptcy cases did not have to confront in the same way.
The Employee Privacy Gap
The earlier cases involving Toysmart, Borders and RadioShack largely centered on consumer information.
Spirit brings employee information into the center of the dispute.
That difference matters because employees provide enormous amounts of information to their employers as part of doing their jobs.
- They use company email.
- They communicate through company messaging systems.
- They create documents.
- They participate in internal meetings.
- They submit employment and payroll information.
- They interact with HR systems.
- They create software and operational records.
- They generate calendars, schedules and other digital records.
Much of that information is created because the employment relationship requires it.
Employees generally do not decide that they are going to create a commercially valuable dataset for a future buyer. They are performing their jobs inside systems controlled by their employer.
Spirit’s bankruptcy raises the question of what happens to that accumulated digital history when the employer disappears.
That is a different question from whether a customer agreed to a privacy policy on a website.
It is a question about the status of employee-generated data itself.
Why This Could Become a Landmark Case
It would be premature to call the Spirit matter a landmark legal decision. The court has not yet approved the sale, and there has not yet been a ruling establishing a new legal rule.
But the case has the characteristics of a dispute that could become important beyond Spirit.
First, the scale is extraordinary.
The proposed dataset covers decades of corporate activity and includes enormous volumes of communications and records.
Second, the proposed buyer is a major AI company.
Google is not buying the airline or continuing to operate its business. The stated purpose of the transaction is to acquire business data and other assets that can be useful for improving products and AI models.
Third, the dispute concerns employee information rather than only conventional customer databases.
Fourth, the transaction is taking place within a bankruptcy proceeding, where the company’s assets are being monetized after the company ceased operations.
And fifth, the parties are confronting a technological reality that did not exist when the earliest bankruptcy privacy cases were decided: enormous collections of workplace data can now be processed by AI systems at a scale and speed that were previously impossible.
Those facts do not establish that the Spirit transaction violates privacy law.
They do establish why the transaction deserves close scrutiny.
New Question for AI Age
The Toysmart, Borders and RadioShack cases established a basic tension.
On one side is the bankruptcy system’s need to identify and monetize assets. On the other is the privacy interest of people whose information became part of those assets.
Spirit adds a third dimension.
What happens when the asset is not simply a customer list, but decades of organizational memory, employee records and workplace communications, and the potential buyer is an AI company? That is the question that makes this case different.
The issue is not simply whether Google can see someone’s name. It is whether information created during an employment relationship can acquire a new economic purpose after that relationship and after the employer itself has disappeared.
And the Spirit Airlines Google data sale raises an even broader question for every company that collects and stores employee information:
- Who owns the economic value of an employee’s digital history?
- Who controls it after the employee leaves?
- What happens to it when the company fails?
And what privacy obligations follow the information when it becomes an asset in a bankruptcy sale?
There are already legal mechanisms for dealing with personal information in bankruptcy. The Toysmart, Borders and RadioShack cases demonstrate that clearly. But those cases primarily dealt with consumer information and privacy promises made by businesses to their customers.
The Spirit dispute asks whether those existing concepts are sufficient for an era in which employees generate enormous quantities of interconnected digital information every day and AI companies have a commercial incentive to acquire and process such data.
The court’s eventual decision will not answer every one of those questions. But it could help define where the boundaries lie. That is why the Spirit case matters.
It is not because bankruptcy sales of personal data are unprecedented. They are not. It is because the nature of the asset has changed.
For decades, companies have accumulated employee data as a byproduct of running their businesses. Now that data itself can have a measurable market value.
And once employee data becomes a valuable asset, the question of who gets to control it after a company’s collapse is no longer theoretical. It becomes a question for bankruptcy courts, privacy regulators, employers, employees and, increasingly, the AI industry.
The Spirit case may ultimately be remembered not for the $10 million price attached to the data, but for the question it forced into the open:
When a company disappears, does its employees’ digital history disappear with it, remain protected, or become something that can be sold?
That question is bigger than Spirit Airlines. And it is only beginning to be tested.
Reference:
- Association of Flight Attendants-CWA: Spirit bankruptcy data-sale objection
AFA: Spirit Bankruptcy: Objection to Sale of Your Data - FTC: Toysmart settlement
FTC: Settlement With Bankrupt Website Toysmart.com - FTC: Toysmart bankruptcy court order on sale of customer information
FTC: In re Toysmart.com, LLC, Stipulation and Order - FTC: Toysmart case file
FTC: Toysmart.com case documents - FTC: Borders bankruptcy and customer information
FTC: Protection for Personal Customer Information in Borders Bankruptcy - FTC: RadioShack bankruptcy and customer information
FTC: Protection for RadioShack Consumers’ Personal Information
- Reuters: Google to buy Spirit Airlines business data for $10 million
Reuters: Google to buy Spirit Airlines business data for $10 million - Reuters: Court delays hearing after union objects
Reuters: US court delays hearing on Google’s purchase of Spirit Airlines data - Axios: Google wins bankruptcy auction for Spirit’s emails, chats and documents
Axios: Google wins bankruptcy auction for Spirit Airlines data - Ars Technica: Flight attendants’ privacy concerns
Ars Technica: Flight attendants object to Google’s Spirit data purchase - Wall Street Journal: Spirit flight attendants fight Google’s data bid
Wall Street Journal: Spirit Flight Attendants Fight Google’s Data Bid for AI
Leave a Reply