DoorDash Starts Paying Couriers for Data to Train AI and Robotics

DoorDash has begun paying some couriers not only for deliveries, but also for digital tasks: for example, for uploading short videos and other content to improve AI models. In this way, the platform is turning its field network into a source of data for machine learning and robotics.

How It Works

According to Bloomberg, in some regions DoorDash is offering couriers paid tasks that are not directly related to delivering an order to the customer. This involves short video clips and other digital actions that help improve AI and robotics models. This is an important shift: previously, platforms like DoorDash primarily monetized logistics, but now they are also beginning to monetize access to a distributed network of people who are constantly "in the field" and can collect useful data from real-world environments.

This approach fits well with the gig platform economy. DoorDash already has an app, a micropayment system, a geography of performers, and a task distribution mechanism. Adding another layer of work on top of delivery—digital tasks—is simpler for the company than building a separate contractor network for data collection.

For couriers, this looks like an additional way to earn between orders, and for the platform itself, it's an expansion of the business toward AI infrastructure without a radical restructuring of its core service.

Why DoorDash Needs Data

AI and robotics models face constraints not only in computing but also in data quality. They need fresh examples from the real world: what urban environments look like, how lighting conditions change, how objects behave in everyday scenarios. Videos and other digital tasks completed by couriers can provide exactly this kind of material—not sterile laboratory data, but live and varied. For systems that need to navigate the physical world, such information is particularly valuable.

For DoorDash, this scheme offers several advantages:

• the company receives a stream of new data without creating a separate field team;
• collection can be quickly scaled through the already existing network of performers;
• couriers can be offered additional microtasks between deliveries;
• the logistics platform gains another growth path in the AI direction.

At the market level, this is also an indicative move. The Bloomberg article directly states that DoorDash is following competitors who have already found new ways to use gig workers in the AI boom. This suggests that demand for flexible human work in the AI industry does not disappear even against the backdrop of automation. Quite the opposite: the more actively companies develop models, the more they need people to collect, verify, and replenish datasets.

New Gig Economy

The DoorDash story shows how the very role of platform labor is changing. A courier is no longer just transporting a package from point A to point B. He can become a participant in a digital pipeline that supports model training. In such a scheme, physical work and data work gradually merge: the same performer delivers food today and tomorrow in parallel helps improve computer vision algorithms or robotic systems.

This turn has practical questions. If such a model begins to scale up, the market will need to discuss how to pay for such tasks, where the line is drawn between delivery and digital side work, what requirements are placed on content quality, and how transparent platforms are in explaining exactly what the collected materials are used for. For the company, it's an inexpensive way to build up its data pipeline, but for workers, it's an even more fragmented, atomized form of labor, where each additional step becomes a separate microtask.

What This Means

DoorDash demonstrates that the next stage of the AI economy is being built not only around models and chips, but also around access to real people and offline data. Companies with large performer networks are gaining a new asset: the ability to quickly collect material for training AI. For the labor market, this is a signal that gig work is increasingly being embedded not just in delivery, but in servicing the AI industry itself.