DoorDash is rolling out a standalone Tasks app that pays its couriers to capture videos, photos, and audio designed to train artificial intelligence and robotic systems. The company says pay is shown upfront and calibrated to the effort and complexity of each assignment, positioning the program as a flexible way for Dashers to earn beyond deliveries while contributing to next‑generation AI.
Assignments range from filming everyday activities to recording multilingual speech samples. The goal is to gather high‑quality, real‑world data that helps models better interpret environments, objects, and human instructions—capabilities essential to logistics, robotics, and computer vision at scale.
What Dashers Will Do And Why It Matters For AI
DoorDash confirmed that tasks include hands‑on videos—such as washing a set of dishes while wearing a body camera—and structured speech recordings in other languages. Bloomberg reported that this original media will be used to evaluate both DoorDash’s in‑house AI and models from partners in retail, insurance, hospitality, and technology.
Not all assignments are standalone. New digital tasks will also surface inside the main Dasher app, like photographing real menu items to improve listings or capturing hotel entrances to guide drivers to accurate drop‑off points. DoorDash’s collaboration with Waymo—where couriers are paid to close self‑driving car doors—also appears in the task list, signaling how the company blends human feedback with autonomous systems in the field.
Scale is the key advantage. DoorDash cites more than 8 million Dashers across the U.S., a workforce that can collect data nearly anywhere, quickly, and with granular instructions. That reach is difficult for traditional data vendors to match and could accelerate development of embodied AI that needs egocentric, real‑world context rather than synthetic or studio footage.
Availability And Payout Transparency For Dashers
The standalone Tasks app and the in‑app tasks are launching in select U.S. locations, excluding California, New York City, Seattle, and Colorado. DoorDash says it plans to expand to more task types and countries. While specific rates were not disclosed, the company emphasizes upfront pricing and effort‑based pay, an approach meant to help workers decide quickly whether an assignment is worth their time.
The exclusions hint at the complex interplay between local labor and privacy laws and new AI data practices. Markets like NYC and Seattle have unique pay rules for app‑based workers, while California and Colorado enforce robust data‑privacy regimes. Clear, localized compliance will be crucial as the program grows.
Why AI Needs This Kind Of Data From The Real World
Robotics and multimodal AI falter when trained solely on web images or lab datasets. They need egocentric video, ambient audio, and diverse speech to handle messy, real‑world conditions—dim lighting in a restaurant kitchen, reflective surfaces on dishes, or the variations of signage at a hotel entrance. Research communities have underscored this need, with initiatives like the Ego4D project demonstrating how first‑person video improves object understanding and manipulation planning.
For DoorDash, better perception translates into practical wins: more accurate navigation cues, fewer failed deliveries, smarter restaurant recommendations, and eventually safer human‑robot collaboration. For partners, these same datasets can tune risk models in insurance, guest‑experience tools in hospitality, and shelf or menu recognition in retail.
Gig Work Meets The AI Data Economy At Scale
DoorDash is not alone. Uber has piloted small jobs that pay drivers to upload images for AI training, while data vendors such as Appen and Scale AI have long relied on flexible workforces to annotate and collect media. The twist here is proximity: Dashers are already embedded in the environments AI needs to learn, from storefronts and kitchens to apartment lobbies.
Success will hinge on task design and quality control. Clear prompts, short task durations, and straightforward validation can keep effective hourly pay competitive. If assignments require body cameras or multi‑step setups, safety guidance and equipment standards will matter, along with concise instructions to reduce rejections and rework.
Privacy And Governance Questions For Real-World Data
DoorDash will face scrutiny over how it handles personal data captured in the wild. Best practices include automatic face and license‑plate blurring, location‑metadata controls, strict retention limits, and clear rights around how long and where media is used. Frameworks from the NIST AI Risk Management Framework and the Partnership on AI’s data‑sourcing guidelines offer useful guardrails.
Dashers will also want clarity on consent flows, the ability to opt out of certain task categories, and what happens if third‑party partners reuse their media. Transparent policies and easy‑to‑read task briefs can build trust and sustain supply as demand for real‑world data ramps up.
What To Watch Next As DoorDash Expands Its Tasks App
Key signals to monitor include average payout per minute for common tasks, the acceptance and completion rates across regions, and how quickly DoorDash expands beyond the initial markets. Expect competitive responses from other logistics platforms and retailers with large field networks, as companies race to secure high‑value, first‑person datasets that are still scarce.
If DoorDash can pair rigorous data governance with reliable earnings and consistently high media quality, Tasks could become a durable second income stream for couriers—and a strategic pipeline of real‑world intelligence for the AI systems increasingly orchestrating commerce and mobility.
