Google’s $10 Million Spirit Data Deal Gains Union Privacy Safeguards

Page 2 of the October 8 congressional letter published by Representative Steven Horsford’s office requests worker-data safeguards before transfer; it predates the November union agreement. Source: Office of Representative Steven Horsford.

Worker protections follow passenger exclusions

The stakes were unusually broad workplace records, created or supplied because people worked for Spirit. An October 8 congressional statement, citing court findings, put the proposed archive at approximately 100 million emails and 500 million Microsoft Teams messages, alongside employment records. Those are figures for the earlier proposal, not a count of what Google will ultimately receive.

The statement accompanied a letter from 121 lawmakers seeking safeguards before transfer. It warned that removing direct identifiers did not necessarily prevent identification through patterns and connections across records.

Passenger privacy had already received separate attention. The November 10 account said a privacy ombudsman had recommended approval the previous month after excluding passenger personal information from 97 million Spirit customers. It also recounted Google’s October statement that it did not want personal information and that such information would be excluded or de-identified by an independent third party before receipt. That was Google’s stated position, rather than a finding that processing had been completed.

The employee agreement adds exclusions and verification to that earlier position. AFA’s detailed list of materials outside the sale, which it identifies as part of the bill of sale, includes medical and disability data, drug and alcohol testing records, disciplinary and grievance files, bank and payroll records, identification documents, biometric data and recordings of flight attendants. It also excludes multiple aviation-safety datasets, including confidential safety-program materials, fatigue reports and flight-crew operational communications.

A report on the revised terms describes the exclusions as encompassing a decade of payroll records and employee tax forms, along with timecard information dating back to 2012.

For data remaining in the sale, AFA says an independent verification agent will review the de-identified records and will be selected in consultation with the unions. Google must keep the data confidential, process it securely and refrain from trying to re-identify it.

The revised proposal names Tonic AI Inc. as the company to perform de-identification, according to an account of a notice filed in the U.S. Bankruptcy Court for the Southern District of New York. Preparing the data and independently reviewing the result are separate steps; the accounts of the agreement do not establish that either has been completed.

Approval and the remaining archive are the next tests

The November 10 account placed court approval in the following week. Resolving union objections therefore clears an obstacle to the transaction rather than completing it.

The account of the revised court proposal also describes a right for Google to withdraw if search-term exclusions reduce the volume it is buying. The accessible text ends before stating the full condition, leaving the threshold unclear.

The next decisions concern court approval and whether the archive left after exclusions and privacy processing satisfies the sale’s terms. The $10 million bid prices the proposed transaction; neither the earlier message counts nor the separate passenger figure establishes the size or value per record of the final dataset.

Amazon’s Anthropic investment could be worth $272 billion at IPO

Amazon’s shares and convertible notes could be worth $271.8 billion at the top of Anthropic’s proposed IPO range. Existing accounting gains, a 12% ownership cap and Anthropic’s computing commitments shape what that paper value means.