There's something quietly unsettling about Google's $10 million purchase of bankrupt Spirit Airlines' corporate data and it isn't the price tag.
For a tech giant, the sum is little more than a rounding error, but it buys a treasure trove of institutional memory: 100 million emails, 500 million Microsoft Teams chats, 7.2 billion records on competitors' flights, 7.5 billion passenger transaction records dating to 2008 and more than 175,000 employee records going back to 1986.
Thrown in for good measure is around 30 million lines of code, development metadata, and software models and algorithms.
Google insists this is a sanitised transaction. A spokesperson said the firm will not receive any personal information and that any data received will be rigorously scrubbed of personally identifiable information by a third party before receipt.
The deal explicitly excludes Spirit's 97.5 million passenger profiles and frequent flyer records. On paper, that sounds like a reasonable firewall.
But the objections raised by Spirit's former workforce deserve serious attention.
Sara Nelson of the Association of Flight Attendants-CWA said her union is filing a court objection to Google's attempt to buy data that has no business being sold, arguing that de-identification requirements in the sale agreement do not address confidential records.
Anyone who ever emailed a colleague at Spirit, filed an expense report or raised a grievance through internal channels is now, in effect, training material for a trillion dollar company's language models. De-identified is not the same as consequence free.
There is also the industry angle and this is where LCRAB readers should sit up.
Google is not buying Spirit's data out of nostalgia for low-cost carriers.
The truth is that AI firms are hunting for messy, specialised, real-world operational data that just isn't available on the open internet.
Aviation, with its pricing curves, scheduling logic, HR processes and fraud detection workflows, is exactly the kind of white collar domain these companies now want to model and eventually automate.
That should give every regional and low-cost carrier pause.
If a defunct airline's back office correspondence and pricing data can be auctioned off as training fodder, what happens to the data of carriers still flying?
Contracts, HR systems and IT vendor agreements were not written with AI training rights in mind and bankruptcy proceedings have shown how quickly that data can become a saleable asset once a company folds.
The competing bid from Mercor.io, and a later higher offer from Micro1, only underline how contested this market has become. A federal judge still needs to approve the Google sale.
Whatever the outcome, the case has opened a debate the industry cannot avoid. Who actually owns the digital exhaust of running an airline and what happens to it when the airline is gone?
