Can You Earn Extra Cash by Recording Videos of Folding Clothes? DoorDash Tasks Allows Delivery Drivers to Side Hustle and Train AI, Sparking Heated Discussion

American English translation:

U.S. delivery giant DoorDash today launched a new Tasks app, allowing delivery workers to capture daily activities or record voice content to help AI models and robots understand the real world. This move opens a new revenue stream for the gig economy and reflects a trend where many industries are collecting data from the real world to train their own models or sell externally.

(Pokémon players train 30 billion photos to build an “AI world model,” aiding the delivery robot industry)

DoorDash Launches Tasks: Delivery workers can earn extra income from daily chores

DoorDash announced the launch of a standalone app called Tasks, enabling millions of its delivery workers to take on additional tasks for compensation. These tasks include various daily activities such as folding clothes, doing dishes, making beds, and even trimming plants. Payments depend on the complexity and time required, typically ranging from a few dozen dollars.

Additionally, the platform offers voice recording tasks, such as asking users to hold natural conversations in specific languages to assist in training voice models.

DoorDash stated that this data will be used to improve AI and robots’ understanding of the physical world, further applied in automation and intelligent systems development. The company also noted that Tasks is currently a small-scale pilot, with plans to gradually expand the types of tasks and application scenarios in the future.

From text and images to real-world data: AI training needs moving toward physical behaviors

In recent years, sources of AI training data have expanded from text and images to more complex physical behavior data. DoorDash’s new initiative exemplifies this trend. By collecting human actions, operational processes, and language interactions in real environments, AI models can more accurately simulate human behavior—such as learning how to properly load utensils into a dishwasher or understanding how items are arranged in a home.

TechCrunch reported that the video and audio data collected by DoorDash will not only be used for internal AI models but may also be shared with retail, insurance, hospitality, and tech industry partners for application and testing, further increasing the value of the data.

Gig economy as the ideal training ground for AI data: Uber and others follow suit

DoorDash is not the only company integrating gig workers into AI training. Uber previously experimented with a similar program, allowing drivers to upload photos and recordings for extra income; Instawork has recruited workers to wear head-mounted devices to record household chores. Even robot companies like Sunday Robotics collect human operation data through “motion capture gloves” to train household robots.

This trend shows that gig platforms are gradually becoming “data collection networks,” leveraging their large user bases to quickly gather diverse real-world data across different regions.

Community discussion: AI training enters the era of “human behavior commercialization”

Following the exposure of the Tasks program, related discussions have emerged online. A producer from Bankless believes that collecting daily behavior data through monetary incentives could lead to a trend where gig workers become data producers, which is crucial for developing automation and robotics applications.

As DoorDash continues to collaborate with autonomous driving companies to deploy contactless delivery services, balancing efficiency improvements, labor shifts, and data usage will become a key concern for the industry and society.

This article: “Just film yourself folding clothes to earn extra cash? DoorDash Tasks sparks debate over gig workers training AI” first appeared on Chain News ABMedia.