The BaseResearchFederal Reserve Bank of Chicago
Forecasting the Economic Effects of AI
In a Chicago Fed working paper, researchers surveyed 69 economists, 52 AI industry and policy professionals, 38 superforecasters and 401 members of the public on AI's effects on the US economy. Unconditional forecasts stay near historical trends; under a rapid-progress scenario, economists forecast labor force participation falling to 55% by 2050. These are forecasts, not measurements.
Open the source ↗Top story of the week
Why it matters
The same thing, for different readers.
What it means for you
Surveyed economists and AI experts mostly expect modest change in jobs and growth, but widely uncertain outcomes if AI becomes far more capable; these are forecasts, not measurements.
For lawmakers
The authors note that policy cannot plan for the median: if AI progress is rapid, outcomes for growth and participation spread widely, including a possible deep fall in participation.
For researchers
Scenario-conditional forecasts split disagreement about AI's pace from disagreement about its effects, finding the second larger; the authors judge scenario ambiguity a limited caveat.
The question it raises
Economists and the public want different remedies although they forecast broadly similar outcomes, so the choice between retraining, a job guarantee and basic income is not settled by forecasts alone.
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“The median respondent in each group expects substantial advances in AI capabilities by 2030, small declines in labor force participation consistent with demographic shifts, and an annual GDP growth rate of 2.5%, which exceeds both the typical medium-run (2.0%) and long-run (1.7%) baseline forecasts from government agencies and private-sector forecasters.”
Found word for word in the source
“Conditional on a “rapid” AI progress scenario … experts forecast substantial, though not historically unprecedented, economic shifts: annualized GDP growth rising to around 4% and the labor force participation rate falling from its current level of 62% to 55% by 2050, with roughly half of that decline—equivalent to around 10 million lost jobs—attributable to AI.”
Found word for word in the source
“A variance decomposition suggests that expert disagreement about these effects is driven primarily by different beliefs about the economic effects of highly capable AI systems rather than by disagreement about the pace of AI progress.”
Found word for word in the source
“Economists strongly favor targeted policy interventions such as AI-focused worker retraining (71.8% support) over broad structural interventions like job guarantees (13.7%) or universal basic income (37.4%), whereas the general public supports both targeted and broad interventions.”
Found word for word in the source
“We also note that worlds with rapid AI progress likely differ from worlds with slow progress in ways that extend beyond AI capabilities themselves, so this gap should be understood as the estimated effect of moving between scenarios that are focused on AI rather than a cleanly identified causal effect of AI capabilities alone.”
Found word for word in the source
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- A survey of forecasts, run online from October 2025 to February 2026 and taking about 8 hours on average: 69 economists, 27 AI industry professionals and 25 AI policy professionals (grouped as 52 AI experts), 38 superforecasters and 401 members of the general public. They forecast US GDP growth, productivity, labor force participation, unemployment, wealth inequality and other outcomes for 2030 and 2050, both overall and conditional on slow, moderate and rapid AI-progress scenarios for 2030, gave their probability of each scenario, and rated six policy proposals. Only the economists' answers are reweighted, for non-response.
- What it found
- Every group expects substantial AI progress by 2030 (economists give 61.4% on average to the moderate or rapid scenario), yet unconditional forecasts stay near recent trends: median GDP growth of 2.4% to 2.5% to 2030 in every group, and, for economists, labor force participation falling from 62.6% (January 2025) to 61.0% in 2030 and 58.3% in 2050. If AI progress is rapid, economists forecast GDP growth of 3.5% in 2045-49 (superforecasters 4.0%, public 4.5%, AI experts 5.3%) and, for 2050, participation of 55.0%, a labor share of income of 45.0% (55.48% in 2025) and 80.0% of wealth held by the top 10% of households (71.2% in 2023), with a much wider range of outcomes; the authors attribute roughly half of the participation decline, around 10 million jobs, to AI. A variance decomposition finds expert disagreement driven mainly by differing beliefs about the economic effects of highly capable AI, not about the pace of progress. Economists support retraining (71.8%) and modernized unemployment insurance (62.3%) more than basic income (37.4%) or a job guarantee (13.7%); the public supports a job guarantee (57.1%) and favors basic income more than it opposes it (47.9% to 30.3%).
- What it does not show
- Forecasts, not measurements: nothing here records what AI has done to jobs so far. The scenarios describe AI capabilities in 2030, not adoption. The authors note that two people choosing "rapid" may have pictured different capabilities, that forecasters may have over-anchored to the historical data shown in the survey, and that the scenarios may miss the most extreme beliefs about capabilities. The 10-million-jobs figure compares the rapid and slow scenarios and, the authors say, is not a cleanly identified causal effect of AI. Only economists were reweighted for non-response; the other groups are small (52 AI experts, 38 superforecasters) and describe the recruited samples, not wider populations. Of 4,866 experts contacted, 147 completed the survey before filtering, and the economist sample skews younger, less experienced, more male and more European than the sampling frame. The public sample came from an online platform. Funded by Open Philanthropy (now Coefficient Giving). A working paper, not edited by the Chicago Fed; its views do not necessarily reflect those of the Chicago Fed or the Federal Reserve System.
- Status
- A working paper: not yet peer reviewed. The paper reports no independent replication. It compares with an abridged version of the survey run in a separate panel (LEAP, February 2026): experts there gave the rapid scenario a higher probability, while forecasts conditional on each scenario differed less.
The record
- Questions
- What happens to work, income and the economy? · How fast is AI advancing, and where is it heading?
- People
- Ezra Karger (Wrote it, Federal Reserve Bank of Chicago) · Otto Kuusela (Wrote it, Forecasting Research Institute) · Jason Abaluck (Wrote it, Yale School of Management) · Kevin Bryan (Wrote it, University of Toronto) · Basil Halperin (Wrote it, University of Virginia) · Todd Jones (Wrote it, Mississippi State University) · Connacher Murphy (Wrote it, Stanford University) · Phil Trammell (Wrote it, Stanford University) · Matt Reynolds (Wrote it, Forecasting Research Institute) · Dan Mayland (Wrote it, Forecasting Research Institute) · Ria Viswanathan (Wrote it, Forecasting Research Institute) · Ananaya Mittal (Wrote it, Forecasting Research Institute) · Rebecca Ceppas de Castro (Wrote it, Forecasting Research Institute) · Josh Rosenberg (Wrote it, Forecasting Research Institute) · Philip E. Tetlock (Wrote it, University of Pennsylvania)
- What we read
- We read the whole source.
- Found by
- A research pass: work, income and the economy, gathered October 13, 2026, via
research-work(tier 1: the source’s own channel or an official record) - Cite this
- Federal Reserve Bank of Chicago, “Forecasting the Economic Effects of AI”, March 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/chicago-fed-2026-03-forecasting-economic-effects-ai
Our notesRead the main text in full (introduction to discussion); the appendices (about 8,000 lines of tables, question wording and example rationales) were skimmed, not read in full. Fifteen authors; the quotations are the paper's collective words, attributed to the first-listed author (Ezra Karger, a joint first author with Otto Kuusela), as the Base does for multi-author papers. Check before relying on the 10 million jobs: the abstract and body say roughly half of the participation decline is attributable to AI, but the footnote's own arithmetic gives about 60% (4.2 of 7 percentage points, about 11 million jobs) and rounds it to 10 million; the abstract starts from 62% (February 2026) while the body also cites 62.6% (January 2025). The abstract says the median respondent in every group expects 2.5% GDP growth; the body gives medians of 2.4% to 2.5% for 2030 and up to 3.0% (AI experts) for 2050, which is what the evidence above states. The abstract's rapid-scenario GDP growth of "around 4%" blends groups (economists 3.5%, superforecasters 4.0%, public 4.5%, AI experts 5.3% for 2045-49). The research pass's description also calls this NBER Working Paper 35046; the PDF does not say so. Open Philanthropy is named as funder on the first page. importance 3 is our suggestion: the broadest source in this pass on what informed groups expect and where they disagree.