DoorDash wants your burrito guy to train the AI replacing you

DoorDash just looked at its couriers and said: “Congrats, you’re also the camera crew for the robot apocalypse.”

They launched “Tasks,” which pays gig workers to film short videos that get used to train AI models. Not deliver the food. Film the footage. Like Uber Eats, but make it Black Mirror.

DoorDash’s pitch is basically “extra earning opportunities.”

Because this isn’t just labor. It’s labeled data. The kind of stuff that turns into “we trained our system on millions of real-world examples” in a keynote, right before a guy in a hoodie announces your job is now an API.

And here’s the spicy part: once your video helps train an AI that prints money later, do you get royalties? Equity? A cut? Or just the same thing gig work always offers—$12 today and a soothing app notification about “flexibility” while your knees dissolve.

DoorDash isn’t alone. Every big tech platform is quietly realizing the real product isn’t burgers or rides or packages—it’s humans generating high-quality training data while pretending they’re just “working.”

Meanwhile, the ownership question is sitting there like a dead rat in the kitchen: who owns the footage, the labels, the outcomes? Because if DoorDash owns it, they’re not just dispatching couriers—they’re building a private dataset moat out of people who can’t afford a flat tire.

This lands on you too, by the way. Today it’s “couriers filming Tasks.” Tomorrow it’s your workplace asking you to “help improve our tools” by recording yourself doing your job, perfectly, for the machine.