Google agreed to pay $10 million for a massive collection of Spirit Airlines’ internal business data, including employee emails, Microsoft Teams chats, spreadsheets, calendars and operational information.
The companies say customer information and personally identifiable information will be excluded, with the data de-identified by a third party before Google receives it.
Even with those safeguards, the unusual sale raises privacy questions for former Spirit employees whose workplace communications could ultimately be used to help develop and train Google’s AI systems.
Google is preparing to spend $10 million on an unusual asset from bankrupt Spirit Airlines: its data.
The technology giant won a bankruptcy auction for a vast collection of the airline’s internal digital records, including roughly 100 million emails and 500 million Microsoft Teams chats, according to reports on court filings. The package also includes documents, spreadsheets, calendar information, software and data related to areas such as marketing, productivity and airline operations.
Google has said the information could be used to improve its products and train artificial intelligence models. The transaction still requires bankruptcy court approval, with a hearing scheduled for Aug. 19.
The deal illustrates how valuable private corporate information is becoming as technology companies seek new sources of data for AI development. But it also raises an important question: What happens to the privacy of the people whose emails and workplace conversations are part of the dataset?
What data is Google getting?
This isn't primarily a database of passenger records.
The assets reportedly include Spirit employees’ emails, Teams conversations, calendars and spreadsheets, along with business and operational information. Google outbid AI data company Mercor, which offered $7.5 million for the assets.
That distinction matters for consumers. Reports on the proposed transaction say customer information and personally identifiable information will not be included in the data Google ultimately receives. A third party is expected to de-identify the material before the transfer is completed.
"We acquired part of an enterprise dataset from Spirit Airlines, which can be helpful in improving our products and AI models," a Google spokesperson said in a statement.
"We will not receive any personal information from this dataset. Any data we receive will be rigorously scrubbed of any personally identifiable information by a third party before receipt."
That means travelers shouldn't assume Google is purchasing a database containing their names, credit card numbers or individual booking histories. Based on the publicly reported terms of the deal, those types of customer records are supposed to be excluded.
So, are there privacy concerns?
Potentially — but they are different from the most obvious concern that Google might suddenly receive millions of Spirit passengers’ personal details.
Complicated privacy issue
The more complicated privacy issue involves employees and other people who may appear in Spirit's internal communications.
Workplace emails and chat messages can contain names, opinions, personnel discussions, conversations with vendors and other information that wasn't necessarily written with the expectation that it would someday become training material for another company's AI systems.
De-identification should substantially reduce that risk if it is done effectively. But removing obvious identifiers such as names and email addresses doesn't necessarily eliminate every possibility that a person could be identified from context, particularly across an enormous collection of interconnected communications.
There is also a broader question of expectations. Employees generally understand that workplace communications belong to their employer and may be retained or reviewed. Selling hundreds of millions of those communications as an asset in bankruptcy for AI development is a considerably less familiar use.
The scale makes the issue particularly significant: reported figures put the collection at about 100 million emails and 500 million Teams conversations.
Consumers have reason to pay attention
Spirit has faced privacy litigation before, although it is unrelated to Google's proposed purchase.
In May, a federal appeals court addressed a lawsuit alleging that Spirit used third-party "session replay" software capable of recording website visitors' interactions. The court upheld dismissal because the plaintiffs had not adequately established a concrete privacy injury, while modifying the dismissal to be without prejudice.
That case does not establish that Google's data purchase creates a privacy violation. It does, however, demonstrate how questions about what companies collect, how information is used and what constitutes a legally recognizable privacy injury can become complicated quickly.
A new kind of asset in bankruptcy
Perhaps the biggest significance of Google's purchase is what it says about the value of corporate data itself.
Bankrupt companies traditionally sell aircraft, real estate, intellectual property and other tangible or clearly defined assets to repay creditors. The Spirit auction suggests years of internal emails, chats and operational records can also carry a multimillion-dollar price tag — particularly when AI companies are looking for large collections of real-world information unavailable on the public internet.
Google's $10 million bid also wasn't uncontested. Mercor's competing $7.5 million offer demonstrates that more than one AI company saw substantial value in the dataset.