Uber has initiated a pilot program in the United States that allows its drivers and couriers to perform microtasks to train artificial intelligence models, an effort aimed at establishing the company’s freelance workforce as a competitor to existing AI data-training platforms.
The new program was announced as part of a series of updates designed to create what Uber CEO Dara Khosrowshahi described as “the best platform for flexible work” during an event in Washington, D.C. The central component of this initiative is the introduction of paid digital tasks that drivers and couriers can complete through the Uber app. These microtasks are designed to generate, annotate, and verify data used to train and refine AI systems.
The prompts presented to participants will vary, covering a range of data collection types. For instance, a driver might be asked to record audio clips of themselves speaking in their native language or a local dialect, which helps train voice recognition and natural language processing models. Other tasks could involve capturing and uploading images of specific objects, such as different types of cars, or submitting documents written in particular languages. One example provided by the company is a task to upload a photograph of a menu written in Spanish, a single submission that could earn the participant a payment of one dollar.
This strategic move positions Uber’s extensive global network of independent contractors as a direct challenge to established entities in the AI data services market, such as Scale AI and Amazon’s Mechanical Turk platform. These businesses are foundational to the generative AI industry, providing the human-powered data labeling and annotation necessary to train large-scale models. The work typically involves humans reviewing, categorizing, and correcting the vast datasets that AI systems learn from. A significant portion of this labor is sourced from outside the U.S. at a lower cost, and it is widely considered a critical element in the advancement of sophisticated AI capabilities. By integrating these tasks into its existing driver and courier app, Uber is leveraging its workforce to enter this specialized market.
Uber’s involvement in using its independent contractors for AI-related work is not entirely new. The company’s AI Solutions Group has previously detailed its use of “human-in-the-loop” processes. This methodology combines the cognitive abilities and expertise of human workers with the speed and scale of machine automation to improve AI model performance and accuracy. Further signaling its ambitions in this sector, Uber recently acquired Segments.ai, a Belgian startup specializing in data annotation tools, to bolster its growing data-labeling business. The microtask pilot itself is an expansion of a similar program that was previously made available to Uber drivers in India, who were given the opportunity to earn small additional amounts by responding to prompts within the app. The company is now extending this model to its U.S. workforce.
The reception of these microtasks among U.S. drivers and couriers remains an open question. Many workers on the platform have expressed ongoing concerns regarding their earnings, citing what they describe as low pay resulting from the company’s high commission, or take rate, on fares from rides and deliveries. Uber’s business model is predicated on classifying its drivers and couriers as independent contractors, which means they are considered to be in business for themselves. This classification exempts the company from providing traditional employment benefits such as overtime pay, minimum wage guarantees, and contributions to health insurance. Conversely, some drivers contend that the level of control Uber exerts over their work through its algorithm—dictating fares, routes, and performance standards—is more characteristic of an employment relationship than an independent contractor arrangement.
In addition to the AI training pilot, Uber announced several other updates focused on improving the daily experience for its drivers and couriers. One significant change is a redesign of the trip-offer cards that appear in the app. These cards display key information about a potential ride or delivery, which the worker must review before accepting or rejecting the request. The updated design provides drivers and couriers with more time to assess the details of an offer before making a decision. For couriers specifically, Uber is also launching a new on-trip interface designed to simplify the process of handling multi-order deliveries. This revamped experience provides clearer, more organized information for each pick-up and drop-off location and includes a new system of alerts intended to help prevent commonly missed items from an order being left behind.
To provide drivers with better information about where to find work, Uber has introduced a new heatmap. This feature is designed to offer greater transparency regarding demand levels across different areas. The map uses a color-coded system to indicate wait times for trips. Red areas signify the shortest wait times, followed by orange and then yellow, which represent progressively longer waits. Purple areas on the heatmap indicate that surge pricing is currently in effect.
The map also displays the average time drivers waited for a trip in a given area, based on recent historical data. A related feature allows drivers who are commuting from one location, such as their home, to a high-demand zone, like a city center, to choose their preferred type of route. They can opt for a route that gets them to their destination as quickly as possible or select an alternate route that is calculated to maximize their potential fares by directing them through areas with likely trip requests along the way.
Uber also detailed changes aimed at enhancing safety on the platform. The company is expanding the availability of its Women Rider Preferences feature, which was first launched in the United States in July of the previous year. The feature is now being rolled out in additional cities, including Baltimore, Minneapolis, Philadelphia, Seattle, Portland, and Washington, D.C. This tool allows drivers and riders who identify as women to set a preference to be matched only with other women. According to Uber, in the markets where this feature has been active, it has been used on over 100 million trips. The company also reported that a quarter of its women drivers activate the feature each week, and more than half of those who use it keep it enabled for over 90 percent of their driving time.
Furthering its efforts to give drivers more control, Uber is introducing a feature that allows them to set a minimum rating for the riders they are willing to accept. Drivers can adjust this minimum rating based on their personal comfort level and can toggle the preference on or off as their situation changes. For example, a driver might choose to set a higher minimum rating for trips late at night and relax that requirement during daytime hours. Uber stated that this tool, when used in conjunction with its existing rider verification measures, helps “give drivers more peace of mind on every trip.”
Finally, Uber addressed long-standing driver complaints about platform deactivations with several updates related to fairness. Drivers have frequently reported feeling powerless when faced with what they perceive as indiscriminate deactivations and have described the appeals process as difficult and time-consuming. In response, Uber stated that it is implementing changes to mitigate these issues. The company provided a direct statement on the new policy: “We know that losing access to Uber can create real challenges, which is why we’re working to make it easier for drivers and couriers to keep earning, even if issues arise. Where we can, we’ll limit access to only specific types of earnings opportunities rather than the full platform. For example, if an alcohol delivery issue is reported, drivers can still take food or rideshare trips. As always, serious violations, including safety issues, may result in losing access to Uber altogether.”
Under the new system, if Uber receives a complaint about a driver from a rider, the platform will now provide the driver with an opportunity to present their side of the story before a final decision on their account status is made. To further promote fairness, Uber will also hold riders accountable for their reports; if a rider is found to have filed false reports against a driver, that rider may face deactivation from the platform.
The company also introduced a Delayed Ride Guarantee, a feature through which drivers can earn additional compensation if a trip takes longer than expected due to customer delays or other extenuating circumstances such as heavy traffic or necessary detours. In a move to increase driver earnings, Uber is also enhancing its tipping reminders for customers by expanding them to iPhone Live Activities, creating more frequent and visible notifications that prompt riders and delivery recipients to tip their driver or courier.
