I Tried DoorDash’s Tasks App and Saw the Bleak Future of AI Gig Work
Summary
Author style: Punchy — this is a first‑hand, hands‑on dispatch from the frontline of AI data labour. The writer tests DoorDash’s Tasks app by filming everyday chores to see how platforms are turning human behaviour into training data.
The article describes the author’s experience using DoorDash’s Tasks app, which pays gig workers to record short videos or perform simple actions (folding laundry, scrambling eggs, walking in a park) so those clips can be used to train AI systems. The app enforces strict framing and submission rules, providing live feedback (beeps, on‑screen prompts) to get the exact footage required. Pay per task is low, the work is repetitive and invasive, and the process highlights growing concerns about the labour, privacy and power dynamics behind the datasets that fuel AI.
Key Points
- DoorDash’s Tasks app recruits gig workers to capture video and other data to train AI models.
- Workers must follow precise instructions (camera framing, hand visibility, repeated takes) enforced by real‑time app feedback.
- Tasks are simple but highly repetitive and can be awkward or intrusive (filming in private spaces, showing personal items).
- Compensation is low relative to the time and effort required, raising questions about exploitation and labour protections.
- The workflow scales the platform’s power: companies can cheaply collect vast amounts of labelled, real‑world data from distributed humans.
- There are privacy and consent risks for workers and for any people or property shown in submitted footage.
- The piece situates Tasks within a broader trend: platforms expanding beyond their original services to monetise and harvest human activity for AI development.
Why should I read this?
Because it’s weirdly compelling and a bit grim — you get a front‑row seat to what daily life looks like when companies ask ordinary people to be AI’s unpaid sensors. If you care about fair pay, privacy, or where the data that trains tomorrow’s AI actually comes from, this is the short, vivid read that saves you time by showing you the practical, human side of the problem.
Context and Relevance
This article matters because it puts a human face on the invisible workforce building AI: rather than abstract contractors in a dataset pipeline, these are app users asked to perform intimate, repetitive tasks for modest pay. It highlights key industry trends — platforms diversifying revenue streams, the commodification of everyday behaviour, and the outsourcing of dataset generation to low‑paid gig workers. Policymakers, labour advocates and anyone tracking AI ethics should note how easily data collection can scale and how poorly regulated that labour currently is.