According to Business Insider, Uber has suddenly terminated AI training contracts for workers on “Project Sandbox,” its partnership with Google, just one month into what was supposed to be a three-month gig. About a dozen contractors, many with PhDs or advanced degrees, were told via email that Google had “communicated a change in their internal priorities.” These workers were earning up to $110 per hour for tasks like annotating photos and evaluating AI-generated answers, with potential monthly earnings around $19,000. Now they’re being let go two months early and are still waiting for their first paychecks, which Uber said could take up to seven weeks to arrive. The company told workers their assignments would conclude immediately and they need to return company-provided laptops.
The gig economy reality check
Here’s the thing about the “future of work” that companies like Uber are selling: it’s incredibly fragile for the workers actually doing the labor. These weren’t random people – Uber specifically recruited PhDs for this work, promising stable three-month contracts. Then poof, gone after four weeks. And workers are stuck waiting for pay while being told to ship back their equipment immediately. One contractor told Business Insider that other companies they’ve worked for actually honor their contract lengths. But when you’re essentially disposable labor, even advanced degrees don’t guarantee job security.
Uber’s bigger AI ambitions
What’s really interesting here is the timing. Uber CEO Dara Khosrowshahi was just talking up the company’s AI training business on their earnings call earlier this month, saying some work requires PhDs and positioning Uber as becoming a broader “platform for work.” Meanwhile, they’re cutting their most qualified workers with zero notice. The company’s stock is up 39% this year, and their core ride-hailing business is posting double-digit growth. So why the sudden pullback on what seemed like a strategic new revenue stream? It suggests that Uber’s AI Solutions division might be more vulnerable to client whims than they’d like to admit.
The AI training industry’s dirty secret
Labeling data for AI has become a massive global business, employing hundreds of thousands of people worldwide. Some do it as a side hustle, others as their main income. But this incident exposes how precarious this work really is, even at the high end. Workers reported that Uber would sometimes cut their hours below the 40 needed to qualify for the top $110 hourly rate. Now they’re dealing with sudden termination and payment delays. Basically, the companies building billion-dollar AI models are relying on a workforce that has zero job security. When your client (in this case, Google) changes priorities, you’re just collateral damage.
What happens now?
Uber says they’re “committed to keeping you in our network” and will reach out with future opportunities. But how much is that worth when they just demonstrated they’ll cut people loose without warning? The workers are now in limbo – waiting for pay, waiting for return instructions for their laptops, and waiting to see if Uber actually follows through on finding them new gigs. For a company trying to position itself as the future of flexible work, this doesn’t exactly inspire confidence. And it makes you wonder – if this is how they treat their most educated contractors, what does that say about their broader gig work platform ambitions?
