DoorDash just turned its courier fleet into an AI training army. The delivery giant quietly launched Tasks, a new app that pays gig workers to film themselves doing everyday activities - from cooking dinner to speaking foreign languages - creating a novel revenue stream that blurs the line between delivery work and data labor. The move signals how platforms are racing to monetize idle worker time while feeding the insatiable appetite for training data that powers modern AI systems.
DoorDash is betting that its army of delivery workers has something more valuable to offer than just dropping off tacos - their everyday experiences captured on camera. The company's new Tasks app transforms couriers into paid data collectors, filming mundane activities that will eventually train the next generation of AI models.
The platform lets workers pick up micro-tasks during downtime between deliveries. Need someone to film themselves chopping vegetables? Record a conversation in Mandarin? Walk through a grocery store aisle? Tasks has a gig for that. Each completed assignment adds a few dollars to a courier's earnings, creating what DoorDash frames as flexible income opportunities.
But this isn't just about giving workers more ways to make money. The launch positions DoorDash squarely in the booming AI training data market, where companies spend billions acquiring the videos, images, and audio clips that teach machine learning models to understand the world. By tapping its existing workforce, DoorDash sidesteps the need to recruit dedicated data collectors or purchase expensive datasets from specialized vendors.
The timing isn't coincidental. AI companies are burning through training data at unprecedented rates, and the quality bar keeps rising. Generic stock footage won't cut it anymore - models need diverse, real-world scenarios captured from multiple angles and contexts. A courier filming their actual morning routine provides authenticity that staged content can't match.
For gig workers, Tasks represents both opportunity and complication. The app promises earnings flexibility, letting couriers choose which tasks to complete and when. Someone waiting for their next delivery ping can quickly film a 30-second clip and pocket a few bucks. But it also raises thorny questions about compensation. If that video helps train an AI model worth millions, is a $5 task payment fair? Who owns the rights to footage of a worker's home, family, or daily life?
