Uber announced at its Only on Uber 2025 conference that it's piloting a program to let drivers complete AI data labeling tasks through the app when they're not driving. The digital microtasks include uploading vehicle photos, recording audio in different languages or accents, and submitting documents like restaurant menus in various languages. The pilot's already testing in India and expanding to the US.

Compensation varies by task complexity—some jobs like menu uploads pay up to $1 each. Drivers see the payment upfront before accepting, similar to how they accept rides. Uber says the tasks aren't related to autonomous vehicle development—they're for external AI clients through Uber's AI Solutions Group.

Uber's competing with Amazon Mechanical Turk, Scale AI, and Upwork for the data labeling market. That market's worth $6.5 billion in 2025 and projected to hit $19.9 billion by 2030. Scale AI does $290 million in annual revenue with a $29 billion valuation.

Uber already has the platform and the workers. Millions of drivers who know how to accept tasks through an app, get paid, move on to the next gig. That's what you need for data labeling—quick tasks, instant payment, massive scale. Uber just needs to add the task matching logic.

The downtime angle makes sense. Drivers wait between rides, especially in slower markets or off-peak hours. Sitting in your car waiting for the next ping? Snap a photo of your vehicle or upload a restaurant menu for $1. Stack a few tasks and you've made $5-10 in idle time. Uber's monetizing driver downtime.

But is this Uber's core competency? They're a mobility and delivery platform. Data labeling is different—you're competing with specialized companies like Scale AI that spent years building client relationships with AI labs, optimizing quality control, and developing domain expertise.

Uber already runs an internal data labeling team called Scaled Solutions for their own rideshare, delivery, and freight operations. They expanded it to serve external clients like Aurora Innovation (self-driving trucks) and Niantic (Pokémon Go). So there's precedent.

The opportunity is undercutting pricing with their existing driver base. Scale AI pays Filipino, Nigerian, and Kenyan workers for data labeling. Uber could pay US drivers competitive rates while charging clients less because they're not building a platform from scratch—they're adding features to existing infrastructure.

Whether it works depends on execution. Can Uber maintain quality when drivers do quick tasks between rides instead of dedicated labeling shifts? Will drivers want this work, or is switching contexts (driving → labeling → driving) too much friction? Can Uber's AI Solutions Group compete with companies that spent years building client trust?

The gig economy creating more opportunities—drivers get more ways to earn without working more hours behind the wheel. You decide when to drive, when to label data, when to do nothing. That's the core value.

The announcement came with other driver changes: women rider preference rolling out to more cities (used on 100 million+ trips), rating preferences so drivers can avoid low-rated passengers, and delayed ride guarantees for longer trips.

These changes came from 60+ crew sessions with drivers. Uber listened and built features. The AI data labeling pilot fits—drivers wanted more earning options during downtime.

Critics will say Uber's exploiting drivers by turning them into data laborers. But drivers opt in, see payment upfront, and choose which tasks to accept. You know exactly what you're earning before you start. That's transparency.

Gig platforms keep expanding. Uber started as ride-hailing, added food delivery, freight, advertising, and now AI data labeling. DoorDash did the same—started with restaurant delivery, expanded to convenience, groceries, and autonomous robots. These platforms are becoming multi-purpose marketplaces.

Uber's got S&P 100 status now, which gives them resources to experiment. Data labeling might work, might not. But testing it shows how gig platforms leverage existing infrastructure to enter adjacent markets. Once you've got millions of users accepting tasks through an app, you can plug in whatever makes economic sense.

If Uber captures 5% of that $6.5 billion market, that's $325 million in annual revenue from a feature bolted onto existing infrastructure. Worth piloting.

Whether this becomes real revenue or a footnote depends on driver adoption and client demand. Low risk, high upside, leveraging existing strengths. That's platform growth.

Source: CNBC