TaskNet: Building Occupational Task Inventories from Worker Interviews at Scale

Jennifer Wang, Andy Haupt, Isaac Robinson, Phil Lewis, Erik Brynjolfsson, Diyi Yang · Stanford Digital Economy Lab · 1 Jan 2026

Read original source

Where this sits in the debate

Will AI destroy more jobs than it creates?

Direction of the finding as read from the abstract by an AI model (AI-written).

Key takeaway AI-written

Stanford's TaskNet builds an occupational task taxonomy directly from worker interviews, with each task statement written to O*NET standards and validated by job analysts and workers. Across 132 interviews in six occupations it yields 202 task statements, 79 of them new to O*NET, while 30 existing O*NET tasks never came up.

The paper in three steps AI-written

Written by an AI agent from the paper's abstract, which is the primary source and is quoted below.

The question

Can interviews with workers be turned, at scale, into an occupational task list that captures how jobs are changing with AI before formal job descriptions catch up?

The approach

The authors built TaskNet from 132 worker interviews, 2,726 task reports and 7,923 ratings from 318 raters across six occupations. Each task statement is written to O*NET standards and validated by job analysts and workers.

The findings

TaskNet holds 202 task statements: 79 are new to O*NET and 123 overlap with it, while 30 existing O*NET tasks never came up in the interviews.

How to read it: dataset-or-tool. Introduces data or methods others build on — its impact shows up in the studies that use it.

Limitations and counterpoints AI-written

The inventory covers six occupations, so it does not yet describe the wider labour market, and the abstract does not report how the tasks relate to AI exposure or use.

Source excerpt

Understanding how AI is reshaping work requires an accurate account of the tasks workers perform. Workers are often the first to see how their jobs are changing, as new responsibilities and tools often appear in everyday work before these changes reach formal descriptions. Interviews with workers can help capture these emerging changes in job tasks. Yet existing methods support standardized ratings across workers, while open-ended conversations with workers remain costly to collect and synthesize at scale. We introduce the Stanford Task Network (TaskNet), an occupational taxonomy grounded directly in workers' accounts of their everyday work. Every task statement is written to O*NET's standards and validated by job analysts and workers. Across 132 worker interviews, 2,726 task reports and 7,923 ratings from 318 raters in six occupations, TaskNet holds 202 task statements: 79 new to O*NET, 123 overlapping with O*NET, while 30 O*NET tasks never came up in interviews.

About this publication

Industry report. An industry or consultancy report — produced outside academic peer review.

Published: 2026 (10 months ago)
Citations: Not recorded for this source