Uber doesn’t just want to move people anymore, it wants to train AI, build robotaxis, and maybe, just maybe, reinvent what “work” looks like in the process.
During Uber’s Q3 earnings call on Tuesday, CEO Dara Khosrowshahi made it clear: the company isn’t content being just a ride-share and delivery empire. It wants to be an AI powerhouse too. Two of the company’s five new strategic pillars are now entirely AI-focused and one of them might directly involve you.
Turning Drivers Into AI Trainers
Uber recently launched a pilot program called Digital Tasks, a feature that lets drivers and couriers earn extra cash by helping train AI models.
Yes, you can now drive people, deliver food, and label datasets all within the same app.
Here’s how it works: drivers complete small “micro-tasks” like uploading images, annotating short clips, recording voice samples in their native language, or rating responses generated by AI systems. It’s currently live in India and the U.S., but Uber says expansion is on the way and it won’t stop with drivers.
“Some of the roles require PhDs, for example, in physics, in order to get the gig done, so to speak,” Khosrowshahi said.
This all falls under Uber AI Solutions, the company’s growing AI data-services business. According to Khosrowshahi, that division is already “landing a ton of customers.”
From “Rides” to “Platform for Work”
At first glance, this might sound like an odd detour for a company built on moving people from point A to point B. But Uber sees a bigger picture: it wants to become a platform for all kinds of work, not just transportation.
“Another way of looking at our platform is that we’re a platform for work, and the first kind of work that we have gone after is transportation,” Khosrowshahi explained. “But we can empower other kinds of work as well, which is what Uber AI Solutions is all about.”
It’s a bold idea, one that could turn Uber into the go-to “everything app” for flexible, on-demand jobs. Whether it’s driving, delivering, or teaching AI to recognize a coffee cup, Uber wants it all happening inside its ecosystem.
The Other AI Bet: Robotaxis
Of course, Uber’s AI ambitions don’t end with human labor. The company is also going all-in on autonomous vehicles, the ultimate AI workhorse.
Uber plans to integrate “human drivers and autonomous vehicles into a single marketplace,” combining the two worlds into one seamless service.
To make that happen, Uber recently announced a massive partnership with Nvidia to build a fleet of 100,000 robotaxis, with production kicking off in 2027.
That’s a huge number, but there’s a catch. Even Khosrowshahi admits robotaxis aren’t profitable yet, and won’t be for “at least a few more years.”
The Safety Question
Then there’s the elephant in the room: safety.
Just last week, a Waymo robotaxi fatally struck a beloved bodega cat in San Francisco, an incident that ignited public outrage and led local officials to propose laws that could let cities ban autonomous vehicles altogether.
And guess what?
Uber is testing robotaxis in San Francisco right now, in partnership with Lucid Motors and Nuro. The timing couldn’t have been worse.
Still, Khosrowshahi remains optimistic, predicting that in ten years, “every single new car sold” will include autonomous capabilities.
“That is a very bright future for the world, because it will make the world safer,” he said.
My Take: Uber’s AI Play Makes Sense – But It’s a Tightrope
I actually think Uber’s approach here is smart, but risky.
On one hand, the Digital Tasks idea is a clever use of Uber’s global workforce. The company already has millions of people logging in daily, often with downtime between rides. Turning that idle time into AI training is pure efficiency. It’s crowdsourced data collection at scale, something OpenAI and Anthropic spend billions trying to manage.
But on the other hand, this move blurs the line between flexibility and exploitation. If Uber starts relying on human gig workers to feed the AI that will eventually replace them (through robotaxis and automation), the optics could turn ugly fast.
And then there’s execution. Building reliable robotaxis and sustainable AI services takes more than just scale, it takes trust, regulation, and a lot of capital. Uber is trying to juggle all three.
If they pull it off, Uber could transform from a ride-share company into one of the most versatile work platforms on Earth. But if they stumble, this could easily look like a company chasing trends instead of perfecting what it already does best.
The Bottom Line
Uber doesn’t just want to take you places anymore, it wants to teach machines, run fleets of driverless cars, and redefine the gig economy itself.
That’s an ambitious goal. Whether it leads to a safer, smarter world or just another overextended tech pivot will depend on one thing: how much of this AI dream can actually make money.