AI companies such as Mercor are paying white-collar workers to produce the training data needed to automate their jobs. In this feature at The Verge, Josh Dzieza speaks with dozens of workers and documents what the labor looks like: low-paying, precarious, exploitative, and designed to extract expertise from entire professions—law, science, writing—before moving on to the next. It’s “the gig economy to the very extreme,” one screenwriter tells Dzieza, and it’s building a future with no place for the people who built these industries. (Subscription may be required.)
If you move fast and get lucky and have the right combination of expertise and stay on the right side of each platform’s unique and mysterious recipe of productivity metrics, you can make decent money. I spoke to a playwright making $10,000 a month, a multitalented chemist who at various points found gigs demonstrating poker and singing for AI. But even then, there is an inescapable awareness of ephemerality because producing training data means working toward your own obsolescence. While the number of people doing data work may continue to rise, any particular gig will last only as long as it takes for the machines to successfully mimic it. It takes years for a human to develop expertise, and sooner or later, they’re going to run out of skills to sell.
More picks from Josh Dzieza
Friend or Faux?
“Millions of people are turning to AI for companionship. They are finding the experience surprisingly meaningful, unexpectedly heartbreaking, and profoundly confusing, leaving them to wonder, ‘Is this real? And does that matter?’”
The Cloud Under the Sea
“The Ocean Link was one of a small number of ships that maintain the subsea cables that carry 99 percent of the world’s data.”
AI Is a Lot of Work
“When AI comes for your job, you may not lose it, but it might become more alien, more isolating, more tedious.”
