Anthropic Model Tests How AI Could Reshape US Economy

Anthropic, the company behind the Claude AI assistant, has released an interactive economic modeling tool designed to help Americans reason about how artificial intelligence might reshape the U.S. economy by 2030. The Econ Scenario Explorer, built by the company’s economics team, projects three distinct futures — ranging from a gentle boost to a profound transformation with unprecedented growth and sharply higher unemployment.

The tool, accompanied by a technical working paper and a survey of nearly 11,000 U.S. adults, does not predict which outcome is most likely. Instead, it invites users to set their own assumptions about AI’s capabilities and adoption and see the economic consequences those assumptions imply, as NPR reported.

A Model Built on Tasks

At its core, the model represents every job in the economy as a bundle of tasks drawn from the U.S. Department of Labor’s O*NET taxonomy. For each task, AI can augment a worker, automate the task, leave it unchanged, or create entirely new tasks. Workers are split into two groups: knowledge or cognitive workers, who made up 62.4% of the workforce in the 2025 Current Population Survey, and all other workers.

According to the Anthropic Econ Scenario Explorer, the answer depends on how capable the technology becomes, how quickly it spreads, and whether it supports or replaces workers. As Anton Korinek, Anthropic’s head of transformative AI economic studies, put it, “If the AI can do amazing things but nobody uses it, then it’s not going to have an economic impact.”

Three Futures for 2030

The model lays out three scenarios, measured in 2030 GDP at 2025 price levels:

• Modest: GDP 1.6% above a no-AI path at $34.1 trillion, with growth of 2.4% a year and unemployment essentially unchanged at 3.9%. AI’s impact here resembles that of the internet.
• Substantial: GDP 8.3% above the no-AI path at $36.3 trillion, with growth of 5.4% a year — faster than the dot-com boom. Cognitive worker unemployment rises to 4.5%, while other wages climb 5.9%.
• Extreme: GDP 32.4% above the no-AI path at $44.4 trillion, with growth of 15.4% a year and the economy doubling every 4.5 years. Cognitive worker unemployment reaches 17.9% and economy-wide unemployment 11.9%. Anthropic says this path would likely require recursively self-improving AI.

In the extreme scenario, the labor share of income falls from 60% to 45.2%, with capital’s share rising to 54.8%. The paper estimates that a transfer of about 9% of GDP — roughly the size of Social Security and Medicare combined — would be needed to hold cognitive workers’ income at its no-AI level, a scale with no precedent.

What the Public Expects

Anthropic surveyed a representative sample of 10,980 U.S. adults via Morning Consult between August 11 and 23, 2026. Running the median respondent’s answers through the model yields outcomes close to the substantial scenario: GDP about 8.6% higher by 2030 and unemployment around 4.6%. Roughly 10% of respondents held views aligned with the extreme scenario.

The survey also captured how the public sees AI today: 53% said AI can already write routine business emails and documents, and 24% said it can already build and maintain a working software product. On the horizon, 39% expect a Nobel-level scientific discovery by 2030, while 40% say that will never happen.

Anthropic co-founder Jack Clark struck a measured tone, telling NPR that the gap between capability and adoption may be wide. “I think the technology will keep developing at a very, very fast and sustained rate but diffusion of the technology will likely be more challenging than people think,” he said. “So it will get really, really good. But it will make its way into the economy more slowly.”

Analysis: Preparation, Not Prediction

Anthropic is careful to frame the explorer as a scenario-planning tool rather than a forecast. The model attaches no probabilities and omits policy responses, business cycles, financial-market disruptions, catastrophic risks, and hyper-capable robotics — which is a key reason the analysis ends at 2030.

The company’s own economists acknowledge those limits. As Unite.AI reported, the model considers only AI’s effects on cognitive tasks and does not follow individual workers, so it captures the costs of job displacement only coarsely. Outside economists including Daron Acemoglu, David Autor, and David Romer reviewed an early draft; Anthropic says they were not asked to endorse its conclusions.

Even the most disruptive scenario carries an optimistic thread. “If you end up with this level of GDP growth, you have moves available to you as a [government] policymaker that are unimaginable today,” Clark told NPR. “Policymakers should get ready to spend.”

The broader labor-market backdrop remains mixed. Stanford Digital Economy Lab research cited in coverage found no economy-wide AI displacement but a roughly 19% employment shortfall among workers ages 22-25 in highly AI-exposed occupations, driven mainly by reduced hiring. Anthropic says the model will inform the research it funds through its Economic Futures program on interventions for labor-market disruption.

What to Watch

Anthropic describes the explorer as Version 1.0 and says it will evolve as evidence develops. The authors write that data observed over the coming years will indicate which scenario the economy is actually in — and that none of the three can yet be ruled out. For policymakers, the tool’s central message is less about any single number than about readiness: the economic consequences of AI may hinge as much on how institutions respond as on how capable the technology becomes.