DoorDash Pays 8 Million Couriers to Film Chores and Train AI Robots in 2026

DoorDash is paying its delivery workers to strap on body cameras and film themselves washing dishes. The footage trains robots to do the same job. That's not a dystopian thought experiment — it's a new app called Tasks, launched this week for up to 8 million couriers across the United States.

The instructions are specific. Wear the camera. Scrub at least five dishes. Hold each clean one steady in frame before moving to the next. Other assignments include loading a dishwasher, folding clothes, and recording unscripted conversations in Spanish. DoorDash says pay is shown upfront and varies by effort and complexity.

What the data actually does

This isn't busywork. Robotics companies need millions of hours of real human movement to teach machines how objects feel, flex, and break. A dishwashing video teaches a robot how to grip a wet plate without dropping it. A folding video teaches it how fabric bunches and resists.

DoorDash already has skin in the robotics game. Last September, it unveiled Dot, its own autonomous delivery robot that travels up to 20 miles per hour. It partnered with Alphabet's Waymo, paying couriers to close robotaxi doors — also listed as a "Task." And it works with Serve Robotics for sidewalk deliveries in Los Angeles.

The submitted footage doesn't just train DoorDash's own models. Bloomberg reported the data also goes to partner firms in retail, insurance, hospitality, and tech. DoorDash is building a data pipeline, not just a delivery network.

The exclusion map tells its own story

Tasks isn't available in California, New York City, Seattle, or Colorado — four jurisdictions with the country's strictest gig worker protections and data privacy laws. DoorDash says it plans to expand later. For now, it's rolling out in places where fewer rules govern what happens to footage filmed inside workers' own homes.

No details have been published on consent frameworks, data retention, or whether workers can delete their clips after submission.

Everyone's doing this now

DoorDash isn't first. Uber launched a similar program in October 2025, letting drivers upload photos and record speech in different languages during downtime. California-based Sunday Robotics ships "Skill Capture Gloves" to thousands of people who record household routines — the data trains its home robot, Memo, to clear tables and load dishwashers.

Meanwhile, a recent Verge investigation found that companies like Mercor are paying white-collar workers — laid-off lawyers and scientists — to produce the exact training data needed to automate their own professions.

The pattern nobody's naming

Here's what connects all of it. The gig economy was supposed to be the fallback — the flexible work you could always get when the "real" jobs dried up. Now it's becoming a one-way assembly line: humans produce the data, machines learn the skill, and the humans become less necessary.

A courier filming themselves washing dishes today is generating training data for a robot that could wash dishes tomorrow. Forbes put it bluntly: couriers "are generating training data for robotic systems, including potentially the autonomous delivery robots that could one day reduce demand for human couriers."

DoorDash calls it "a new way to earn." The question it doesn't answer: what do the earners do when the earning is done?