People concerned about privacy may choose not to place their own personal information in a chatbot, which could, among other risks, use it to train AI tools. Workers, however, generally do not get the same choice over the material they create on the job.
A multibillion-dollar industry now buys and sells workplace files and emails for companies seeking fresh data to improve AI models, and it continues to grow. Data-privacy laws tend to focus on consumer protections and rarely address the rights of the employees who produced that information. That leaves open what happens when doctors’ notes, employee contracts and casual Slack exchanges are fed into an AI system.
A proposed agreement between Google and a union representing an unexpected group of workers, flight attendants, may provide a more lasting answer. The deal would create a process intended to separate the data workers generate from personal information about them. At the time of the report, a judge still had to approve the agreement in the following week.
The proposed settlement, described as the first of its kind, is part of the liquidation of Spirit Airlines, which abruptly ceased operations during 2026. In August 2026, representatives of the budget airline stated in a legal brief that it would sell decades of data to Google for $10 million. Google said it would not receive personal information as part of the transaction.
The Association for Flight Attendants, a union with 55,000 members, argued that Google’s proposed safeguards would not prevent serious privacy violations involving Spirit flight attendants. The proposed transfer covered records containing 1.99 billion emails, 664 million Microsoft Teams messages, and employee travel and training information reaching back to 1986.
Under the revised terms, Spirit would be required to hire Tonic AI, a third-party data de-identification company, to exclude information that could potentially be associated with individual consumers or employees. Tonic AI would also be responsible for ensuring that the data, once de-identified, cannot be relinked to specific people. Google promised to keep the data in that newly de-identified form.
Many sensitive records would be excluded from the sale, including employee medical records, documents connected to employee investigations and labor relations, and information about industry-wide safety management systems.
In early November 2026, a trade organization representing airlines intervened in the sale. It argued that including that information could make employees less willing to report safety issues through voluntary programs, which are widely regarded as critical to the strong safety record of the US aviation industry.
The proposed agreement also gives Google an exit clause: it may abandon the purchase if the data-cleaning process reduces the assets by more than 8 percent. Google did not respond to a request for comment.
Sara Nelson, president of the Association for Flight Attendants, told WIRED she was as surprised as anyone that flight attendants were at the forefront of AI-related workers’ rights issues, a field usually focused on employees who, willingly or not, train their machine replacements.
“It’s very hard to replace a flight attendant—you’d need a high-functioning robot, and that seems to be further off,” she said. “What this shows is that this is an issue even in jobs you think are most protected from AI. This is an issue for everyone.”
Nelson said the union would immediately begin demanding that other airlines include worker-data-protection clauses in their labor agreements. “If you’re not willing to meet these standards for your workers, we’re going to shame you,” she said. The Association for Flight Attendants has also begun meeting with state-level politicians to develop legislation that would put worker data protections into law.
Ari Ezra Waldman, a professor at the UC Irvine School of Law who studies law and technology, called the proposed settlement “a pretty good outcome” for Spirit Airlines employees. In an email to WIRED, he wrote that unions would now know to negotiate privacy-rights protections ahead of future sales of data assets. He also pointed to another, potentially easier alternative: banning the sale of data entirely. “Privacy protections are necessary now more than ever,” he said.
Technology leaders have argued that workers understand their information will train AI models and that doing so will ultimately make their lives better and easier. “You have to think of the upside, the abundance that AI can result in,” one executive told WIRED in November 2026.
Nelson said US workers do not accept that promised upside and are part of a broader electorate that appears to be growing more wary of the technology. She described the vision of AI abundance as “a fantasyland of people in boardrooms with beanbags who wear hoodies.”