TLDRs;
- Uber expands its AI data labeling pilot to the U.S., allowing drivers to earn money from digital microtasks.
- Tasks include photo uploads, voice recordings, and document verification used to train AI systems.
- Pay varies by task complexity, with earnings deposited within 24 hours of completion.
- Uber’s AI Solutions platform connects over 8 million global workers to data labeling clients worldwide.
Uber is blurring the line between ride-hailing and artificial intelligence as it rolls out a new pilot program that allows drivers and couriers to earn money by helping train AI systems.
The company’s AI Task Program, initially tested in India, is now being expanded to the United States with full deployment expected by the end of 2025.
Through the Uber Driver app, workers can now take on short digital “microtasks” such as uploading images, recording voice clips, and verifying documents. The content collected is used to improve the performance of AI models, especially in speech recognition, image classification, and autonomous driving.
Each task is assigned a fixed payment depending on its complexity and estimated completion time. Uber says that once a task is completed and approved, drivers receive payment directly in their app balance within 24 hours, providing a quick and accessible form of side income.
A New Side Hustle
While the initiative offers flexibility, Uber emphasizes that this is not a replacement for steady income. Availability of these digital gigs depends on client demand for data labeling, meaning that task flow can be unpredictable.
Uber’s setup resembles platforms like Amazon Mechanical Turk, where workers perform small online jobs that typically pay anywhere from a few cents to a few dollars. Each job is optional, with drivers able to view estimated time, complexity, and payout before accepting.
However, Uber has not disclosed throughput rates or guaranteed task volumes, which means that drivers could face significant income swings. For many, these microtasks will likely serve as a supplementary income stream, rather than a consistent revenue source.
AI Solutions Becomes Uber’s Quiet Growth Engine
Behind the scenes, this effort is powered by Uber AI Solutions, the company’s enterprise data services platform. It already provides annotated data to major clients like Aurora and Tier IV, both of which are active in the self-driving technology space.
Uber AI Solutions boasts more than 8 million potential contributors across 100+ languages, enabling it to serve AI companies seeking diverse, region-specific datasets. Tasks can include everything from sentiment detection and image annotation to LiDAR data labeling, a crucial component in training autonomous vehicle systems.
By tapping into its massive global driver and courier network, Uber can deliver culturally and linguistically rich data, positioning itself as a strong competitor to established labeling firms like Scale AI. This gives the company a strategic advantage in the rapidly growing AI training market.
The Business of Labeling Intelligence
Data labeling has quietly become one of the most valuable services in the AI ecosystem. According to market estimates, the global data labeling sector was worth nearly US$4 billion in 2024 and is expected to grow to over US$17 billion by 2030.
Uber’s entry into this space marks a diversification from its traditional mobility and delivery business lines. Instead of only connecting riders to drivers, Uber is now connecting AI companies to human intelligence, effectively transforming its driver base into a distributed workforce for the digital age.
While the long-term sustainability of such side work remains uncertain, the experiment reflects Uber’s strategy to evolve beyond transportation and into data infrastructure for artificial intelligence. For now, it’s a glimpse into how gig work may increasingly merge with the AI economy, one task, one dataset, and one driver at a time.

