AIORG-W008 Working papers

Shifting Work Patterns with Generative AI

NBER w33795

working paper abstract and pdf

Recorded claims
2
coded from the inspected source
Evidence classes
2
scholarly-full-text, adverse-or-corrective
Recorded limitations
5
stated, never hidden

From the wave-0 evidence ledger

Recorded claims

Each claim states what the inspected source says, at the recorded location—bounded by its scope and graded by its confidence. Nothing here is a synthesis across works.

Across 66 firms and 7,137 workers followed for six months, approximately 80% of treated workers used the tool; among users, the intervention reduced email time by roughly two hours per week and reduced after-hours work.

confidence: high within study scholarly-full-text

Scope: Knowledge workers in participating firms over six months; time use rather than output quality or profit.

The study detected no consequential change in the quantity or composition of work tasks beyond individual time savings, an organization-level null that cautions against assuming tool rollout alone produces workflow redesign.

confidence: medium-high adverse-or-corrective

Scope: Measured tasks and six-month window; absence of detected redesign is not proof that redesign cannot occur.

Boundaries

Limitations & independence

Recorded at coding time, carried with the work forever. A claim without its limits is not evidence.

Recorded limitations

  • Treatment take-up was incomplete
  • Digital-trace measures may miss offline work
  • Six-month horizon
  • No direct firm financial outcome
  • Author/product conflict disclosed

Source independence

Academic field experiment across firms; some authors are Microsoft employees and the intervention concerns a Microsoft product.