An open, vendor-neutral standard for measuring how AI is actually changing work: task-level substitution, absorption, and creation, published as a comparable time series.
Join the working groupSurveys, anecdotes, and vendor claims. No one measures AI’s effect on work at the task level, with a shared method, over time.
The rate at which tasks shift between people and AI, attributed with a common methodology and published on a fixed cadence.
Open methodology, versioned governance, no commercial dependency. Anyone can implement it. No one owns the number.
A common language for the units of work AI can absorb, aligned to existing occupational data.
Separating AI substitution from macro cycles, attrition, and ordinary automation — the hardest problem, solved in the open.
What counts as telemetry, minimum quality thresholds, and anonymization standards for contributed data.
Public drafts, comment periods, and versioned releases, so the number can be trusted and reproduced.
We are forming the initial working group: economists, workforce and operations leaders, labor researchers, and builders with task-level data. Leave your email and methodology v0.1 will reach you when it opens for comment.