DoorDash has launched a new standalone app called Tasks, aimed at its Dashers (delivery couriers). It allows them to earn extra money by completing short activities that generate real-world data for training AI and robotics models.
TASKS INVOLVE FILMING THEMSELVES PERFORMING EVERYDAY HOUSEHOLD CHORES, SUCH AS: LOADING A DISHWASHER, HAND-WASHING DISHES, FOLDING CLOTHES AND MAKING A BED. OTHER EXAMPLES INCLUDE RECORDING UNSCRIPTED CONVERSATIONS (E.G., IN SPANISH OR OTHER LANGUAGES) OR CAPTURING PHOTOS/VIDEOS FOR VARIOUS PURPOSES.
Pay is shown upfront and varies based on the task’s effort and complexity, some reports mention $5+ for basic chores, up to $20+ for longer audio recordings.
The data collected—primarily original audio and video footage—helps DoorDash evaluate and improve its own in-house AI models, as well as those used by partners in sectors like retail, insurance, hospitality, and technology. This supports broader goals, including making AI/robotics better at understanding the physical world.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
It’s optional and flexible, often done between deliveries or in spare time. The app is initially available in some U.S. markets, with plans to expand task types and potentially geographies. This fits into a growing trend where gig platforms like Uber or Instacart leverage their large contractor networks to supply high-quality, diverse datasets for AI.
Real human actions in everyday settings are valuable for training models on physical manipulation, object recognition, and more. Some online reactions highlight the irony: gig workers are essentially helping train systems that could eventually automate parts of their own jobs.
Others see it as a pragmatic side hustle in the evolving “AI data economy.” DoorDash’s autonomous robots primarily refer to Dot, the company’s in-house developed, purpose-built autonomous delivery robot. Unveiled in September 2025 by DoorDash Labs (their internal R&D team), Dot represents DoorDash’s push into fully autonomous “last-mile” delivery to handle food, groceries, and other local commerce orders without human drivers.
Dot is compact—about one-tenth the size of a typical car, roughly 4’6″ tall, and weighs around 350 lbs. This makes it street-friendly, able to navigate bike lanes, roads, sidewalks, parking lots, and driveways while fitting through most standard doors for direct-to-door delivery.
Payload: It can carry up to 30 lbs of cargo, enough for items like six large pizza boxes or typical food/grocery orders. Speed: Up to 20 mph, allowing it to cover distances faster than traditional sidewalk-only robots. Fully electric for low emissions and sustainability.
Equipped with a “vision-primary” Level 4 (L4) autonomy stack, including multiple cameras (9+), lidar (3+), radar (4+), and other sensors for real-time perception, obstacle avoidance, and safe navigation in mixed urban environments. Designed with a lower weight and size for better safety profiles compared to full-sized vehicles; it prioritizes visibility to pedestrians and other road users.
Dot was built entirely in-house to address gaps in existing autonomous tech, integrating directly with DoorDash’s marketplace, app, and new Autonomous Delivery Platform for seamless order handoff, routing, and scaling. DoorDash began commercial deployment in late 2025, starting with an early access/launch in the greater Phoenix metro area.
By early 2026, it’s expanding: Partnerships and phased rollouts in other markets, including Fremont, California; manufactured locally there via partners like Sonic Manufacturing Technologies, with initial demonstrations in March 2026 and plans for up to 30 autonomous units in select areas after testing.
Production scaling toward hundreds of units in 2026, with broader U.S. expansion expected. It’s part of DoorDash’s multi-modal strategy, which already includes human Dashers, partnerships for sidewalk robots, and even autonomous vehicles like Waymo in some tests. Availability varies by location—customers in supported areas may see Dot as a delivery option in the app, often for neighborhood or short-range orders.
DoorDash emphasizes it as a way to meet growing demand where recruiting enough human drivers is challenging, especially in suburbs. The recent Tasks app ties directly into this by letting Dashers earn extra by filming real-world chores and activities. This generates diverse, high-quality video/audio data to train and improve DoorDash’s in-house AI/robotics models—helping Dot and future systems better understand physical interactions, object manipulation, navigation in homes, and more.
It’s part of DoorDash’s broader goal to commercialize autonomy at scale in 2026, using real human data to make robots safer and more capable. Dot aims to make delivery faster, cheaper, greener, and more reliable while reducing reliance on gig workers for certain routes—though human Dashers remain central to the platform.