Scenarios for our Economic Future

Anthropic’s economics team has put a speculative price tag on AI’s next few years. Its new scenario explorer asks users to choose how capable AI gets, how widely it spreads, and how much autonomy it has, then shows what the US economy could look like in 2030. The model is built from a technical report called Economic Scenarios for Transformative AI, and it’s framed less as a forecast than as a way to think through possible futures.

The range is wide. In one modest case, US GDP rises 1.6% by 2030 to $34.1 trillion. In a substantial case, GDP is up 8.3% to $36.3 trillion. In the extreme case, output jumps 32.4% to $44.4 trillion. But the headline number hides the awkward part: the richer the economy gets, the more the gains can shift away from workers and toward capital.

Anthropic also tested what ordinary Americans think AI will do. In August, it surveyed more than 10,000 people, and the typical answers lined up with the “substantial change” scenario: GDP 10% higher than it would be without AI, with unemployment around 5%. About 10% of respondents pointed to the most extreme path. That’s not consensus; it’s a spread of expectations, which is exactly the point of the tool.

The labor story gets messier in the more aggressive scenarios. Knowledge workers take the hit first. Some would have to move into jobs like electrician or nurse, which the model treats as less exposed to AI. That kind of switch is slow, and in the extreme case it can leave people unemployed for longer. Wages for knowledge workers are flat in the substantial scenario and fall by more than 10% in the extreme one, while pay outside knowledge work rises.

Anthropic’s own model says the future isn’t fixed, and it also admits what it leaves out: policy responses, business cycles, financial disruptions, even catastrophic risks. Still, the core message is hard to miss. AI may make the pie bigger, but it doesn’t promise a fairer slice without help.