Essentials: productivity
Will AI actually boost productivity and growth? The key papers, numbers, people and newsletters, sized to the time you have. The full question page. Picks are chosen for their weight in the debate and span its sides; they are refreshed every two weeks, last on 29 Oct 2026. The reasons are AI-written and evidence-cited; no rankings or scores.
Five papers, the numbers with their exact quotes, five people and three sources.
Papers to read
- Experimental evidence on the productivity effects of generative artificial intelligenceShakked Noy and Whitney Zhang · Science · 2023 · Evidence for the claim
A preregistered experiment with 453 professionals: ChatGPT cut writing time by 40% and raised quality by 18%.
A preregistered experiment assigning 453 college-educated professionals writing tasks found that access to ChatGPT cut average completion time by 40% and raised output quality by 18%, with the gap between weaker and stronger writers narrowing. Two-thirds of exposed workers reported using ChatGPT in their actual jobs weeks later, making this one of the earliest rigorous demonstrations that generative AI both raises output and reduces within-occupation inequality for a common white-collar task.
- Generative AI at WorkErik Brynjolfsson et al. · The Quarterly Journal of Economics · 2025 · Mixed evidence
Gains measured in a real workplace: 15% more issues resolved per hour, large for novices, small for the most experienced.
Studying the staggered rollout of a generative AI assistant to 5,172 customer-support agents, finds it raised productivity (issues resolved per hour) by 15% on average, with the biggest gains for less experienced and lower-skilled workers, who saw both faster and higher-quality output; the most experienced agents saw little speed gain and slightly lower quality. Because the assistant appears to transfer top performers' know-how to novices, the paper is widely cited as evidence that generative AI can compress skill gaps within an occupation rather than simply replacing workers.
- The simple macroeconomics of AIDaron Acemoğlu · Economic Policy · 2024 · Evidence against the claim
Why task-level gains may add up to little: at most 0.66% more total factor productivity over ten years.
Using a task-based model and Hulten's theorem, estimates that AI's macroeconomic effects over the next decade will be 'non-trivial but modest' — no more than a 0.66% increase in total factor productivity, and likely under 0.53% once harder-to-learn tasks are accounted for — far below headline projections from industry reports. Acemoglu also argues AI could widen the gap between capital and labor income even without worsening cross-group wage inequality, making this one of the most-cited skeptical counterweights in the AI-growth debate.
- Miracle or Myth? Assessing the macroeconomic productivity gains from Artificial IntelligenceFrancesco Filippucci et al. · OECD · 2024 · Mixed evidence
The OECD's general-equilibrium estimate: 0.25 to 0.6 points more TFP growth a year over ten years, between the low and high ends.
This OECD paper combines micro-level performance gains, exposure of activities to AI and expected adoption rates in a multi-sector general-equilibrium model with input-output links. It estimates AI will add 0.25-0.6 percentage points to annual aggregate TFP growth (0.4-0.9 points for labour productivity) over ten years.
- How AI Agents Reshape Knowledge Work: Autonomy, Efficiency, and ScopeJeremy Yang et al. · arXiv · 2026 · Evidence for the claim
Recent production data on AI agents doing knowledge work end to end, a step beyond chat assistants.
Of these 3 papers: 1 with evidence for the claim, 1 with evidence against, 1 mixed.
Of these 5 papers: 2 with evidence for the claim, 1 with evidence against, 2 mixed.
Of these 5 papers: 2 with evidence for the claim, 1 with evidence against, 2 mixed.
Figures to know
- +0.66%TFP gain over 10 years (upper bound), 10 years
Daron Acemoğlu, 2024 · Productivity“no more than a 0.66% increase in total factor productivity (TFP) over 10 years”
- +0.65 pp a year (range 0.4 to 0.9)Extra annual labour productivity growth, 10 years
Francesco Filippucci et al., 2024 · Productivity“(0.4-0.9 pp. for labour productivity)”
- +1.5 pp a yearExtra annual labour productivity growth, 10 years
Joseph Briggs and Devesh Kodnani, 2023 · Productivity“~+1.5pp/yr US labor productivity”
- −40%Time on professional writing tasks with ChatGPT
Shakked Noy and Whitney Zhang, 2023 · Productivity“The average time taken decreased by 40%”
- +15%Customer-support issues resolved per hour with an AI assistant
Erik Brynjolfsson et al., 2025 · Productivity“as measured by issues resolved per hour, by 15% on average”
Every number is quoted from its source and linked to the paper; estimates differ in method and horizon, so they are not directly comparable.
People to follow
- Erik Brynjolfsson agreesEconomist, Stanford University
Co-author of the customer-support study and the productivity J-curve, which argues early gains are undercounted.
Key work: Race Against the Machine: How the Digital Revolution is Accelerating Innovation, Driving Productivity, and Irreversibly Transforming Employment and the Economy; The Productivity J-Curve: How Intangibles Complement General Purpose Technologies
- Daron Acemoglu leans disagreeEconomist, Massachusetts Institute of Technology
Author of the task-based estimate that puts the ten-year productivity gain at no more than 0.66%.
Key work: Robots and Jobs: Evidence from US Labor Markets; The Race between Man and Machine: Implications of Technology for Growth, Factor Shares, and Employment
- Philippe Aghion it's complicatedEconomist; research on innovation and growth
Growth economist whose estimates of AI's effect on annual productivity growth sit above Acemoglu's.
Key work: Competition and innovation: an inverted U relationship; Carbon Taxes, Path Dependency, and Directed Technical Change: Evidence from the Auto Industry
- John Van Reenen it's complicatedRonald Coase School Professor, London School of Economics
LSE economist who directs a programme on innovation and the diffusion of technology across firms.
Key work: The Fall of the Labor Share and the Rise of Superstar Firms*; Are Ideas Getting Harder to Find?
- Anton Korinek agreesEconomist, University of Virginia
Models scenarios in which AI raises growth far above past rates, and what has to hold for that to happen.
Key work: Artificial Intelligence and Its Implications for Income Distribution and Unemployment; Excessive Volatility in Capital Flows: A Pigouvian Taxation Approach
Of these 3 people: 1 agrees, 1 it's complicated, 1 leans disagree. Documented, cited positions only.
Of these 5 people: 2 agrees, 2 it's complicated, 1 leans disagree. Documented, cited positions only.
Of these 5 people: 2 agrees, 2 it's complicated, 1 leans disagree. Documented, cited positions only.
Newsletters and sources
- Justified Posteriors (Andrey Fradkin)Newsletter · 7 posts in this guide
Discusses new studies of AI and productivity as they appear.
- Ghosts of Electricity (Alex Imas)Newsletter · 2 posts in this guide
An economist's newsletter comparing AI with past general-purpose technologies such as electricity.
- Paul KrugmanNewsletter · 5 posts in this guide
Regular posts on whether AI shows up in growth and productivity statistics.
- Anthropic Economic IndexAnthropic · dataset · outside this guide
Recurring data on which tasks people use Claude for, published by Anthropic and not peer reviewed.
- OECD AI Policy ObservatoryOECD · dataset · outside this guide
Home of the OECD's AI productivity papers and cross-country data.