Uber is launching a pilot program that lets US drivers and couriers earn additional income by performing “microtasks” that train AI models. The work includes recording audio, capturing images, and submitting documents in various languages—tasks like uploading car photos, recording speech in local dialects, or photographing Spanish-language menus for as much as one dollar per task.
The move positions Uber’s massive freelance workforce to compete with established AI training platforms like Scale AI and Amazon’s Mechanical Turk. It’s part of a broader announcement from CEO Dara Khosrowshahi about building “the best platform for flexible work,” alongside redesigned trip offer cards, improved heatmaps, expanded safety features for women drivers, and changes aimed at fairer deactivation policies.
Entering the AI Data Labeling Market
Uber isn’t starting from scratch with AI training work. The company has used independent contractors for what it calls “human-in-the-loop” processes that combine human judgment with machine automation, according to its AI Solutions Group. Uber recently acquired Belgian startup Segments.ai to expand its data-labeling business, and drivers in India have already been performing similar microtasks through the app.
The digital tasks pilot brings this work to US drivers, potentially creating a significant new labor pool for AI model training. Data labeling and annotation—work that involves humans reviewing, categorizing, and correcting data that feeds machine learning systems—remains essential for developing powerful AI models despite automation advances. Most of this work currently happens through low-cost labor outside the United States.
What makes Uber’s approach different is the existing infrastructure and workforce. The company already has millions of drivers and couriers using its app daily, complete with payment systems, identity verification, and task distribution mechanisms. Adding AI training tasks leverages that infrastructure without building new platforms from scratch.
The potential scale matters too. Uber operates globally with millions of active drivers and couriers. If even a fraction participate in microtasks during downtime between rides or deliveries, that creates substantial capacity for AI training work—particularly for tasks requiring diverse languages, locations, or cultural contexts that offshore labor pools may not provide.
Compensation Raises Familiar Questions
Whether drivers embrace microtasks remains uncertain given persistent complaints about low pay from Uber’s existing work. The company takes significant commissions on rides and deliveries, leaving drivers with smaller portions of what riders pay. Adding tasks that pay as little as one dollar each for work like photographing menus could feel insulting rather than helpful to workers already frustrated with compensation levels.
The fundamental employment classification issue looms over microtasks just as it does over ride-sharing and delivery work. Uber classifies drivers as independent contractors, arguing they’re in business for themselves and therefore not entitled to traditional employment benefits like minimum wage protections, overtime pay, or health insurance. Many drivers contend that Uber’s algorithm controls their work so thoroughly that they’re effectively employees regardless of legal classification.
Adding AI training tasks doesn’t change this dynamic—it simply creates another category of work performed by independent contractors who lack the protections that employees receive. The microtasks themselves mirror the structure of existing gig economy AI training platforms, where workers perform small digital tasks for minimal pay without employment benefits or guaranteed minimum earnings.
The timing feels particularly fraught. Drivers have spent years advocating for better pay, more transparency, and stronger protections. Introducing microtasks that pay a dollar or less for individual tasks while fundamental compensation and classification issues remain unresolved could be read as tone-deaf or as an attempt to monetize driver downtime rather than addressing core concerns about their primary work.
Redesigned Trip Offer Cards Provide More Information
Beyond microtasks, Uber is making several changes to how drivers interact with the app. Redesigned offer cards—what drivers see before accepting or rejecting trip requests—now provide more time and information for decision-making. The extended viewing window gives drivers additional seconds to evaluate requests before the opportunity disappears.
The company is also rolling out improved multi-order delivery experiences for couriers, simplifying the process with clearer pickup and drop-off details plus alerts for commonly forgotten items. These changes address practical pain points that affect daily work quality and earnings potential.
A new heatmap visualization helps drivers identify high-demand areas through color coding. Red zones indicate shortest wait times, followed by orange and yellow. Purple areas show where surge pricing is active, and the map displays average wait times based on recent data. Drivers commuting from home to busy areas can now choose between routes that get them there fastest or routes that maximize fare opportunities along the way.
These improvements demonstrate Uber listening to driver feedback about needing better tools for optimizing their time and earnings. The heatmap particularly addresses the information asymmetry where Uber’s algorithm knows demand patterns but drivers must guess where to position themselves for the best opportunities.
Women Driver Safety Features Expand
Uber is extending its Women Rider Preferences feature to additional US cities including Baltimore, Minneapolis, Philadelphia, Seattle, Portland, and Washington, DC. The feature, which launched last July, allows women drivers to set preferences so they only receive women riders, and enables women riders to request women drivers.
In markets where it’s available, women have used the feature on over 100 million trips. A quarter of women drivers activate it weekly, and more than half keep it enabled for over 90% of their trips. These usage statistics suggest the feature addresses genuine safety concerns that women drivers experience.
Uber is also introducing dynamic minimum rating controls that let drivers set rating thresholds for accepting riders based on time of day or personal comfort levels. Drivers can enable higher rating requirements for late-night work when safety concerns peak, then relax them during daylight hours when risks feel lower.
Combined with consumer verification for riders, these tools theoretically give drivers more control over who enters their vehicles. However, they also highlight the ongoing safety challenges of work that involves bringing strangers into personal vehicles or entering strangers’ homes for deliveries.
Addressing Deactivation Complaints with Partial Improvements
Driver deactivations have generated persistent complaints about arbitrary decisions and difficult appeals processes. Uber acknowledges these concerns and promises to “make it easier for drivers and couriers to keep earning, even if issues arise,” though the company isn’t eliminating deactivations entirely.
The new approach applies partial restrictions rather than full platform deactivation when possible. If an alcohol delivery issue is reported, for example, drivers can still accept food delivery or rideshare trips. Only serious violations, particularly safety issues, result in complete platform access loss.
When riders complain about drivers, Uber now says it will allow drivers to tell their side before making deactivation decisions. The company also warns that riders who file false reports risk deactivation themselves—a policy meant to reduce weaponized complaints where riders make false accusations to avoid paying or to retaliate for perceived slights.
These changes represent incremental improvements rather than fundamental restructuring of the deactivation system. Drivers still face the reality that their ability to earn depends on Uber’s discretion, with appeals processes that remain opaque and outcomes that feel arbitrary to many workers.
Additional Fairness and Earnings Updates
Uber is introducing a Delayed Ride Guarantee where drivers earn additional compensation when trips get delayed due to customer actions or extenuating circumstances like traffic or necessary detours. This addresses situations where drivers lose time and opportunity without corresponding pay increases.
The company is also expanding tipping reminders for riders, including integration with iPhone Live Activities that creates more frequent nudges about tipping drivers or couriers. While tipping shouldn’t substitute for adequate base pay, riders often forget or decline to tip, and reminders demonstrably increase tip rates.
Building the Ultimate Flexible Work Platform
Khosrowshahi’s vision of “the best platform for flexible work” encompasses all these announcements—from AI training microtasks to improved trip cards, safety features, and deactivation policies. The framework positions Uber as more than a ride-sharing and delivery company, instead as a comprehensive platform for people seeking flexible earning opportunities.
This vision faces tensions inherent to gig economy platforms. Flexibility appeals to workers who want control over their schedules, but that flexibility often comes at the cost of economic security, benefits, and protections that traditional employment provides. The more Uber expands earning opportunities within its ecosystem, the more it reinforces the independent contractor model that keeps drivers outside standard employment protections.
The AI training microtasks particularly embody this tension. They offer additional earning opportunities, but at rates that reflect the same dynamics that make existing work feel inadequately compensated. They provide flexibility to choose tasks and timing, but without minimum earnings guarantees or the benefits that employees receive.
Whether these announcements collectively make Uber “the best platform for flexible work” depends on who’s evaluating and what criteria they’re using. For drivers already satisfied with the platform who want additional earning options, microtasks and improved tools represent genuine improvements. For drivers frustrated with fundamental compensation and classification issues, these changes may feel like rearranging deck chairs while avoiding the structural problems they’ve been highlighting for years.
The AI training pilot will reveal whether Uber’s driver base views microtasks as valuable supplementary income or as another exploitative use of their time and labor. The answer likely varies by individual circumstances, local markets, and how the tasks actually pay relative to the effort required—details that will become clearer as the pilot expands and more drivers try the work.