Media

DoorDash turns gig workers into retail data collectors

Shelf photos and AI training tasks spread to Uber and Instacart, platforms sell surveillance without hiring auditors

Images

DoorDash now lets users train AI or photograph store shelves. 
                            
                              CARLO ALLEGRI/Reuters DoorDash now lets users train AI or photograph store shelves.  CARLO ALLEGRI/Reuters businessinsider.com

DoorDash has begun offering gig workers tasks that look less like delivery and more like field data collection: photographing store shelves, gathering information that can be used for inventory management, retail operations, and training AI systems, Business Insider reports. Uber and Instacart are testing similar “micro-gigs,” extending the app-based labor model from moving goods and people to capturing structured observations about the physical world.

The shift solves a problem for companies building AI and logistics products: data about real stores is expensive to collect with employees, and messy to obtain through formal partnerships. A platform already sitting on a large, on-demand labor pool can turn spare time into a sensor network. Instead of hiring auditors or merchandisers, the company can pay per task, per photo set, or per completed checklist—pricing the world as a stream of datapoints.

For workers, the pitch is flexibility and incremental income. The tradeoff is that the worker becomes responsible for the frictions that normal employment absorbs: being challenged by store staff, navigating rules about photography, and dealing with the consequences of incorrect or incomplete submissions. When the job is “take pictures of this aisle,” the worker is the one standing in a shop explaining why they are documenting shelves.

This also changes the privacy boundary of gig work. Delivery already exposes workers to addresses and routines; in-store data work adds images, metadata, and potentially identifiable details about customers, staff, and security layouts. The platform can keep the contractual relationship at arm’s length while still accumulating a dataset that is durable, reusable, and difficult for the worker to audit or retract once submitted.

The second-order effect is labor-market segmentation. Delivery and ride-hailing are paid for by end customers in visible fees; data collection is typically purchased by retailers, brands, or internal product teams, with the cost buried in supply-chain budgets. That makes the work feel “free” to the consumer while creating a parallel market in which surveillance-adjacent tasks compete on price, not on working conditions.

Business Insider describes the trend as gig work expanding beyond its original categories. The more concrete description is that platforms are turning contingent workers into roaming compliance and data contractors—without giving them the authority, training, or protection that usually comes with that role.

In DoorDash’s case, the new tasks sit alongside food delivery in the same app. The worker still gets rated, timed, and paid per job; only the payload has changed.