Anthropic Maps Three AI Economic Scenarios to 2030

Anthropic released research outlining three scenarios for how artificial intelligence could affect the U.S. economy by 2030. The scenarios range from modest productivity gains with limited labor disruption to an extreme outcome with large output gains and high unemployment.

Anthropic’s economics team built each scenario from assumptions about how rapidly AI capabilities improve, how widely firms adopt the technology, and how much productivity rises on tasks AI touches. The report says the scenarios are not forecasts; Anton Korinek, an economist at Anthropic, wrote, “The scenarios are not predictions, and we attach no probabilities to them; their purpose is to make the consequences of different assumptions comparable.” He added that the next year or two should provide early signals on capability growth and diffusion.

In the modest scenario, AI touches about 4% of economic tasks by 2030 and raises productivity on those tasks by roughly 35%. Instances are split evenly between automation and augmentation, and the model assumes that for every two tasks automated one new task is created and performed by cognitive workers. Under these assumptions, GDP in 2030 would be 1.6% above a no-AI path, annual growth would rise to about 2.4% from 2.0%, average wages would be 0.7% higher, and unemployment would be near 3.9%.

The substantial-change scenario assumes faster diffusion and larger productivity gains. By 2030, AI would affect 12% of tasks, increase productivity on those tasks by about 57%, and automate roughly three-quarters of affected instances. The model assumes that for every four tasks automated one new cognitive task appears. GDP would be 8.3% above the no-AI path by 2030 and annual growth could reach 5.4%. The report anticipates notable shifts in employment across occupations and industries.

The extreme scenario models rapid capability gains and broad adoption. AI would affect 30% of tasks by 2030, more than double productivity on those tasks, and automate 90% of affected instances. The scenario removes the assumption that automation creates enough new cognitive work to offset displaced jobs. The result is GDP 32% above the no-AI path by 2030 and 40% higher than the level in mid-2026, with annual GDP growth of 15.4%. The report estimates average wages about 9.7% above the no-AI path, a shift of roughly 15% of GDP from labor compensation to returns on capital, and overall unemployment near 11.9%.

Anthropic surveyed more than 10,000 Americans in August about expectations for AI adoption and the ease of finding new work. The typical respondent’s answers align with the substantial-change scenario, implying GDP about 10% higher by 2030 and an overall unemployment rate near 5%. About 10% of respondents gave answers consistent with the extreme scenario.

The report highlights wide uncertainty in all scenarios because it is hard to predict how fast AI systems will improve and how quickly firms will integrate them. Anthropic notes that AI experts generally expect faster spread than economists. The researchers use historical benchmarks, including 4.7% U.S. GDP growth in 1999 and peak unemployment just under 10% after the 2008 financial crisis, to compare the modeled outcomes with past economic episodes.