Key Highlights

• Anthropic released an interactive model mapping three scenarios for how AI could reshape the US economy by 2030, ranging from modest growth to historic levels of white-collar unemployment.

• In the most benign "modest" scenario, AI accounts for about 4% of tasks economy-wide by 2030, lifting GDP 1.6% above a no-AI baseline with little effect on unemployment.

• In the "extreme" scenario, AI takes over close to a third of economic tasks, with output surging 32% higher even as Unemployment climbs toward 12% and the labor share of national income falls from 60% to 45%.

• The model's authors say the scenarios carry no assigned probabilities and are not predictions, but are intended to help compare the consequences of different underlying assumptions.

A Tool for Mapping Divergent Economic Futures

Anthropic has released an interactive model mapping three possible scenarios for how artificial intelligence could reshape the US economy by 2030, spanning outcomes from modest growth to historically unprecedented levels of white-collar unemployment. The tool, developed by Anthropic's Economics team, draws on a task-based framework that divides the economy into cognitive occupations, including management, professional, sales and office work, and all other occupations. Users can adjust their own assumptions about AI capabilities, adoption rates and productivity gains to generate projected outcomes for GDP, wages, labor share and unemployment.

Three Scenarios, Three Very Different Outcomes

In the most benign "modest" scenario, AI is used for about 4% of tasks across the economy by 2030. Under that path, GDP would end up 1.6% higher than in a world without AI, with employment and unemployment largely unaffected. The middle "substantial" scenario envisions AI handling 12% of tasks, pushing GDP roughly 8.3% higher than a no-AI baseline while nudging unemployment up to around 4.6%, with knowledge-sector jobs declining by approximately 4%.

At the far end sits the "extreme" scenario, which Anthropic says has no parallel in economic history. Under this path, AI takes over close to a third of all economic tasks, output surges 32% beyond the no-AI trajectory, unemployment climbs toward 12%, and white-collar employment drops by more than 20%. In that scenario, the labor share of national income slides from 60% to 45%, meaning that even with the economy roughly one-third larger than it would otherwise be, total wages flowing to workers would be no greater than in an AI-free world.

A Consistent Pattern Across All Three Paths

Regardless of which scenario ultimately unfolds, the model shows AI redistributing income away from labor and toward Capital holders, with the gap between the two widening as AI takes on a larger share of productive activity. A companion technical paper co-authored by Anton Korinek, Charles Jones, Szymon Sacher, Tess Cotter and Peter McCrory underpins the model's methodology.

Not Predictions, but Tools for Planning

The paper's authors emphasize that the scenarios are not predictions and carry no assigned probabilities, but are instead intended to make the consequences of different underlying assumptions directly comparable. "These scenarios are not predetermined, it's not like an inexorable march," Anthropic economist Peter McCrory said. "Part of the value of doing scenario modeling is so that you can do scenario planning."

A separate Anthropic survey of nearly 11,000 US adults found that median public expectations align most closely with the model's substantial scenario, with respondents anticipating GDP roughly 10% higher by 2030 and overall unemployment rising to around 5%.

A More Gradual Economic Rollout Than Some Expect

Anthropic co-founder Jack Clark said he sees no slowdown in the pace of AI development itself, but believes the path from laboratory research to widespread economic use will prove more gradual than many observers currently expect. "It will get really, really good. But it will make its way into the economy more slowly," Clark said.

Acknowledged Limits of the Framework

The model's authors note several notable omissions in its current design. It does not incorporate the effects of economic downturns, disruptions to financial markets, government policy responses, catastrophic AI failure modes, or the possibility that advances in robotics eventually extend automation into physical labor, all factors that could meaningfully alter the scenarios the model currently maps.

A Framework Likely to Inform Ongoing Policy Debate

With AI's economic trajectory still highly uncertain and the model explicitly designed for scenario planning rather than forecasting, the tool is likely to serve as a reference point for policymakers and researchers seeking to compare the implications of different assumptions about AI's pace of adoption in the years ahead.