Economic Scenarios for Transformative AI (Anthropic, September 2026): extreme case puts GDP growth at 15% a year

This Anthropic Institute working paper by Anton Korinek, Charles Jones and colleagues builds a formal model in which AI automates a growing share of cognitive work between 2026 and 2030, lifting productivity while displacing workers who must search for jobs in other occupations. Three scenarios bracket the range. In the modest case AI adds under half a point to GDP growth and a tenth of a point to unemployment. In the extreme case AI performs almost half of today's cognitive work by 2030, GDP growth climbs to 15% a year, the labor share of income drops from 60% to 45%, and nearly one in five cognitive workers is out of work. A survey of more than 10,000 US adults lands closest to the substantial scenario, where GDP is 8% higher by 2030 and cognitive employment falls 4%. For AI founders the paper is a structured way to argue about demand: how fast automation of cognitive tasks is plausible, which occupations absorb displaced workers, and how wages for knowledge work respond.