Forecasting the Economic Effects of AI

Why this paper is being considered

The paper elicits forecasts of AI progress and US economic outcomes from academic economists, AI-company employees, AI policy researchers, accurate forecasters and the public, under slow, moderate and rapid AI scenarios. It matters because many policy and investment decisions rest on contested beliefs about AI's economic effects. An evaluation would focus on sample selection and response rates, scenario definitions, and how the forecasts should be interpreted given the groups disagree.

What an expert evaluation could add: Downstream fiscal modeling and monetary-policy commentary show a potential audience, but do not validate the forecast elicitation. An expert can assess the method now, including uncertainty and conversion from capability to economic outcomes. Evidence and commissioning case. AI-assisted judgment, 2026-10-07; separate from ratings and completed evaluations.

AI-generated criterion ratings and reasoning

Scores come from AI prioritization; the accompanying explanations include AI-assisted source checks. These are provisional judgments, separate from human ratings and commissioned evaluations.

What the paper says

We elicit forecasts of how AI will affect the U.S. economy, comparing the beliefs of five groups: academic economists, employees at AI companies, policy researchers focused on AI, highly accurate forecasters, and the general public. 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. Conditional on a “rapid” AI progress scenario, in which AI systems surpass human performance on many cognitive and physical tasks, 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. 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. These forecasts map onto notably different policy preferences across groups: experts strongly favor targeted measures such as worker retraining, whereas the general public supports both targeted programs and broader interventions, including a job guarantee and universal basic income.

Claims to check

• Median respondents in all five surveyed groups expect substantial AI-driven advances/effects on the U.S. economy.
• Beliefs about AI's economic effects differ systematically across academic economists, AI-company employees, AI policy researchers, accurate forecasters, and the general public.
• The paper characterizes which group's expectations are most optimistic/pessimistic and how forecasts diverge on magnitude and timing of AI's economic impact.

Opus 5.5 re-evaluation (experimental)

Read with care. This is an experimental re-evaluation by a different model (Claude Opus 5.5, 2026-10-01). Across a calibration sample Opus scored about 15 points lower than GPT-5.5, so 15 points are added to compare it with GPT-5.5 scores. The calibration is a small sample (n=44) and is not human-validated, and the two models disagree about the order of papers within the top group. The AI score at the top of this page is the usual evaluation-priority score shifted by the difference between the adjusted Opus score and the earlier holistic score; the synthesis uses that shifted value, and the unshifted score is shown beside it. On the dashboard you can switch the Opus scores off. Read the methods note.

Different reference score. The earlier score for this paper came from an older Opus run (opus), not GPT-5.5, so no +15 offset is applied: the shift is the Opus 5.5 score minus that earlier score, which assumes the two Opus versions share a scale.

Dashboard details and provenance

Full AI dashboard scoring rationale

This is a fresh NBER working paper eliciting and comparing structured forecasts of AI's effects on the U.S. economy across five distinct populations (academic economists, AI-company employees, AI policy researchers, superforecasters, and the general public). It sits squarely in Unjournal's growing 'economic and social impacts of AI' priority area and overlaps with the forecasting/judgment-aggregation cause area — doubly relevant. The value of evaluation is high: expert-elicitation and forecasting-comparison designs are methodologically subtle (question framing, aggregation choices, selection into groups, calibration/scoring), the results are the kind of headline numbers that get cited quickly in AI-policy debates, and as an unpublished NBER working paper it has not yet had independent peer scrutiny. One of the authors (Ezra Karger) is closely tied to the forecasting-research community, which raises both the paper's likely influence and the case for a genuinely independent read. Main concern: it is an elicitation/survey study rather than a causal empirical estimate, so an evaluator needs to judge the design and interpretation of forecasts rather than an identification strategy — this is evaluable but requires reviewers comfortable with survey/forecasting methodology.

AI decision-relevance rationale

Comparative forecasts of AI's macroeconomic effects inform how funders, think tanks, and policymakers weight AI as a priority and how much credence to place on economist vs. insider vs. superforecaster expectations. It bears on AI-governance resource allocation and on the meta-question of whose forecasts to trust, which matters for GovAI, think tanks, and philanthropic AI-policy funders calibrating expectations of labour, growth, and disruption effects.

AI timing assessment

This is a very recent NBER working paper (w35046) with no peer review yet and authors plausibly still seeking feedback, so an independent evaluation is maximally actionable — it could inform revision before publication and provide early public scrutiny of numbers likely to be cited quickly in AI-policy discussions.

Intake, review, and crux connections

Public attention and use

The paper has several institutional homes (NBER, the Chicago Fed working-paper series, the Forecasting Research Institute) and has been referenced in finance commentary. We found no independent critique. The Fed and FRI links are connected to the authors.

What the search did not establish

• No exact-title EA Forum or LessWrong discussion surfaced in the targeted search.
• No independent critique of the sampling or scenario definitions surfaced.
• The most decision-relevant next signal is whether agencies' baseline forecasts incorporate the surveyed beliefs, or whether a follow-up wave shows belief updating.

Targeted search checked 2026-10-01. Search scope: Targeted exact-title searches across the open web, institutional and author pages, news and public social-media results, EA Forum/LessWrong, and policy/white-paper contexts. Evidence records distinguish commissioning or report use from independent discussion, media attention, indexing, and post-publication policy use. A search miss is reported as uncertainty, not proof of absence. Entries with check_depth "quick" rest on roughly one to two searches and should be read as especially uncertain; entries checked on 2026-10-01 were added for the shortlist of papers without human ratings.

Human feedback so far

The privacy-safe aggregate contains 3 current ratings: 2 team and 1 public. The human mean is 86.7/100. The team has not made a final prioritization decision.

This contradicts the estimates obtained by the series of articles on the AI industrial explosion (see, e.g. https://defensesindepth.bio/the-ai-industrial-explosion-part-4-cheap-power/). Ensure that NO politician uses the estimates from the paper.