Artificial intelligence could deliver a sharp boost to economic growth by 2030 while putting significant pressure on workers in cognitive occupations, wiping out jobs, and crushing wages, according to a new working paper from the Anthropic Institute.
The paper, "Economic Scenarios for Transformative AI," models three possible paths for the US economy between 2026 and 2030 - modest, substantial and extreme. The researchers stress that these are scenarios rather than forecasts, with outcomes depending on AI capabilities, adoption, productivity gains, automation and how quickly displaced workers find new jobs.
The most disruptive outcome comes under the 'extreme scenario', in which AI affects roughly half of the tasks performed by cognitive workers by 2030. The model assumes that 90% of affected instances are automated, rather than augmented, while no new cognitive tasks are created to offset the displacement.
Cognitive Employment Could Fall 21.5%
Under this scenario, employment in cognitive occupations is projected to be 21.5% below its mid-2026 level by 2030. The unemployment rate among workers who began in cognitive occupations rises to 17.9%, compared with a normal level of 2.9%.
Economy-wide unemployment also rises substantially, reaching 11.9% in the extreme scenario.
The paper's definition of cognitive employment covers a broad group of occupations, and the 21.5% figure therefore does not mean that one in five jobs across the entire economy disappears.
GDP Rises, But Labour Gets A Smaller Share
The economic gains in this scenario are substantial. GDP is projected to be 32.4% above the no-AI path by 2030, with annual growth reaching 15.4%, compared with 2% in the no-AI baseline.
But the distribution of those gains changes considerably. The labour share of income falls from 60% to 45.2%, while the capital share rises from 40% to 54.8%.
Capital income is projected to be 81.4% above its no-AI path, reflecting the model's assumption that much of the additional economic output accrues to capital.
Cognitive Wages Fall While Other Wages Rise
In the extreme scenario, wages for cognitive occupations are 11.5% below the no-AI path, while wages in all other occupations are 33.6% higher. As cognitive workers are displaced, labour becomes relatively scarcer in other occupations, pushing wages there higher.
Anthropic estimates that the economic gain in the extreme scenario is almost three times the loss suffered by cognitive workers. In theory, it says, a transfer of around 9% of GDP could keep the affected workers' income at its no-AI level. However, the researchers note that transfers on this scale have no precedent following technological displacement.
The extreme scenario is not the outcome implied by the public's median expectations. Anthropic's survey of US adults produced results closer to its substantial scenario, with GDP 8.6% above the no-AI path and cognitive employment down 4.2% by 2030.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.