Over the past few years, I’ve written a number of posts on the potential economic impact of artificial intelligence. One theme that has consistently stood out is the wide divergence of views on how transformative that impact will be. While there is broad agreement that AI capabilities are advancing rapidly, there is far less agreement on what those advances will mean for economic growth, jobs, and productivity.

“Among tech evangelists in Silicon Valley, it has become conventional wisdom that artificial intelligence will rapidly reshape the labor market, for better or worse. Economists, however, have often discussed A.I.’s impact with a skepticism bordering on dismissiveness,” wrote Ben Casselman, chief New York Times economics correspondent, in a recent article, “Economists Once Dismissed the A.I. Job Threat, but Not Anymore.”

“Recently, however, the message from economists has undergone a subtle change,” he added. “Most still do not see much evidence that A.I. is disrupting the job market. But they are starting to take seriously the possibility that it could someday soon. If it does, they are worried that policymakers are not ready to respond.”

The article cites a March 2026 working paper, “Forecasting the Economic Effects of AI,” based on a survey of how AI might affect the U.S. economy over the next 5 and 25 years. The survey, conducted from October 2025 to February 2026, tracked the views of 69 leading economists, 52 AI industry and policy experts, 38 highly accurate superforecasters, and 401 members of the general public.

The survey results were summarized in a Substack essay, “Forecasting the Economic Effects of AI,” by the Forecasting Research Institute (FRI).

Let me summarize the survey’s key findings.

All groups surveyed anticipate substantial AI capability advances by 2030

Respondents were asked to predict the progress of AI by 2030 by choosing among three scenarios:

Slow progress scenario. AI is a capable assisting technology for humans: writing literature reviews at the level of a capable PhD student, handling half of all freelance software-engineering jobs that would take an experienced human a day to complete, topping up your online grocery cart, and physically unloading dishwashers in some homes.

— Economist: 39%; AI experts: 38%; Superforecasters: 45%; General public: 41%.

Moderate progress scenario. AI is an effective collaborator across domains: autonomous lab systems can accelerate advances in areas like solar-cell technology; almost all freelance software-engineering jobs requiring five days of effort are automatable; robots can do dishes as quickly as humans; robo-taxis can drive anywhere humans can.

— Economist: 47%; AI experts: 46%; Superforecasters: 42%; General public: 41%.

Rapid progress scenario. AI systems surpass humans in most cognitive and physical tasks. Autonomous researchers can compress years-long research timelines into months or even days. AI systems outperform freelance software engineers, customer service agents, paralegals, and clerical workers. Models can write Pulitzer-caliber books — and negotiate the resulting contracts. Robots can operate effectively in homes and factories anywhere in the world.

— Economist: 14%; AI experts: 16%; Superforecasters: 13%; General public: 18%.

All groups anticipate substantial AI capability advances. “The average economist assigned a 61% probability to either the moderate or rapid progress scenario. AI experts gave a similar combined probability (62%), while superforecasters (55%) and the general public (59%) were only slightly more pessimistic.”

Despite predicting significant AI progress, economists do not expect major departures from key economic trends

At first glance, these findings may seem contradictory: strong expectations of AI progress, yet relatively modest expectations for economic change. The survey points to several explanations for this apparent disconnect:

• Historical time lags. As we’ve learned over the past two centuries, there is often a significant lag between the emergence of a general-purpose technology — such as steam power, electricity, automobiles, or computers — and its full economic impact on firms, governments, and institutions.
• Uneven diffusion of AI advances. Productivity gains are unlikely to be evenly distributed across sectors, especially where human labor remains a bottleneck. As the report notes, shifts in capital toward compute, data centers, and infrastructure may take time to translate into measurable GDP gains.
• Geopolitical, structural, and demographic headwinds. Trade tensions, global conflicts, climate change, and political instability — as well as aging populations and constrained immigration — could all dampen economic growth.
• Infrastructure bottlenecks. Constraints in energy supply, chip manufacturing, and data center capacity may limit the pace of AI-driven expansion. Power plants and large-scale infrastructure take years to build, and organizations take time to integrate new technologies.
• Risks in rapid progress scenarios. In more extreme cases, rapid AI progress could introduce additional uncertainties, including societal disruption, geopolitical tensions, or even existential risks. Large-scale labor force withdrawal could also negatively affect growth.

Uncertainty is highest in the rapid progress scenario

Not surprisingly, respondents expressed the greatest uncertainty about economic outcomes under the rapid progress scenario — precisely where the stakes for policy design are the highest.

The rapid progress scenario describes a world in which AI systems surpass humans in most cognitive and physical tasks by 2030. Economists predicted substantial changes when asked to forecast key economic indicators in such a rapid growth scenario:

• GDP growth: 2.4% (2025 baseline) → 3.3% (2030) → 3.5% (2050)
• Total factor productivity: 1% → 2% → 2.5%
• Labor force participation: 62.6% → 59% → 55%
• Unemployment rate: 4% → 6% → 6%
• Wealth concentration (top 10%): 71.2% (2025) → 75% (2030) → 80% (2050)

Even modest increases in growth rates can have large long-term effects. A 3.5% annual growth rate would yield a U.S. economy of $54.7 trillion by 2050 — about 25% larger than under a baseline 2.5% growth scenario.

Job retraining support emerges as the top policy response

Survey participants were asked to evaluate six potential policy responses to AI’s economic impact:

• Retraining support: Up to $25,000 per year (for up to two years) in retraining credits for automation-displaced workers;
• Modernized unemployment insurance: 75% of prior salary for up to 18 months;
• Universal basic income: $1,000 per month for every American adult;
• “Manhattan Project” for AI: deploy 0.4% of GDP annually to accelerate AI R&D;
• A compute tax: Tax heavy AI electricity users $50 per MWh above a set threshold;
• Job guarantee program:  Federally funded job to any adult who wants one.

Retraining support emerged as the clear consensus choice across all groups, supported by 71.8% of economists, 69.2% of AI experts, 76.3% of superforecasters, and 78.5% of the general public.

Support for other policies varied significantly. Economists generally favored more targeted, incremental approaches, while the general public was more open to broader interventions. For example, the job guarantee program was supported by 57.1% of the public but only 13.7% of economists.

The “Manhattan Project” for AI had the highest projected impact on GDP, while the job guarantee program had the largest projected effect on labor force participation. Universal basic income and expanded unemployment insurance had a more limited impact on both GDP and labor force participation.

From Forecasting to Preparedness

In the end, Forecasting the Economic Effects of AI does not provide a definitive answer to how AI will affect the economy. Instead, it offers something arguably more valuable: a clearer understanding of the range of possible outcomes —a nd the reasons for disagreement.

If there is a central message, it is that the future of AI remains highly uncertain, even among experts. This uncertainty is not a weakness of the analysis, but a reflection of the complexity of the phenomenon being studied.

As with previous technological transformations, the ultimate impact of AI will depend not only on the technology itself, but also on how individuals, organizations, and institutions respond to it.

In that sense, the challenge we face is not simply to forecast what AI will do to the economy, but to build the institutions and capabilities needed to adapt to a range of possible outcomes. The question is not just what AI will become, but how we will respond to it.

And that, ultimately, may matter more than any forecast.