AI companies are buying real work files and emails to train their models, and the market is bigger than most people realize. Wired reports that data brokers are collecting internal corporate files, and the buying has reached ordinary inboxes. One freelance writer was offered $1,000 for her work on a single story. She said no.
What is happening
Companies such as Handshake, Micro1 and Mercor buy documents, databases and other records. They then supply AI labs, which use them to teach models how real office work looks. Handshake says it licenses professional documents across many fields, including journalism, and buys graduate students' papers and coursework. Other startups, such as Sunset, offer to help failing companies wind down by scrubbing their data and brokering a sale to the labs.
How big is it
Micro1 CEO Ali Ansari says labs, businesses and robotics firms pay about $40 to $50 billion a year for data. Wired could not verify that figure. He predicts it will pass a trillion dollars a year within two to four years. Of the 65 companies in the latest Y Combinator batch, 40 are classified as data startups. Anthropic says it does not buy workflow from live companies. OpenAI did not respond before publication.
The Spirit Airlines sale
The bankrupt airline put its digital assets up for sale. They included 80,000 email accounts, 500 million internal messages, 20.6 million SharePoint files and a million time-card records. Google seemingly won with a $10 million bid. Micro1 then topped it by $2.5 million, and the two sides argued over who won. The sale is now on hold because the flight attendants' union says it involved their private information.
The privacy question
Data companies say they strip out personal details, and Mercor CEO Brendan Foody says this can be verified. Critics are not convinced. Mayu Tobin-Miyaji of the Electronic Privacy Information Center says emails and chat messages are loosely organized, so private details are harder to find. Workers may have written about doctor's appointments or family problems. Investor Gordon Ritter says standards for this kind of data buying have not been set.
What to do
- Ask your employer what happens to company emails and files in a sale or shutdown.
- Ask whether staff messages would be included and how personal details would be removed.
- Treat work chat and email as something that could be read later, and keep private matters out of it.
It is not yet clear how the Spirit Airlines dispute will end, or whether anonymizing can fully protect workers.
The takeawayWork emails and files now have a market value, so ask your employer what it would do with them in a sale or shutdown.