Anthropic maps three AI futures for 2030; the most extreme could upend the economy

Anthropic researchers have outlined three potential AI futures for the economy between now and 2030. These scenarios are "modest," "substantial," and "extreme," each detailing the impact of AI on jobs, unemployment, and GDP growth.

In the "modest" scenario, AI adds less than half a point to GDP by 2030, growing the economy by just 0.5%, and raises unemployment by just a tenth of a point. This change is steady and gradual, similar to the internet's impact.

The "substantial" scenario predicts AI will be capable of doing half of all knowledge work independently by 2030. This would result in an economy growing at twice its normal rate, but wages for knowledge workers would not rise. The shift would be comparable to the internet or railroad's impact, though reallocation could be costly.

The "extreme" scenario envisions AI becoming more productive than humans on the majority of knowledge work tasks, doing all of them autonomously. This would create an "unprecedented economy" driven by recursively self-improving AI, with GDP growth skyrocketing to 15% per year. However, nearly one in five cognitive workers would be unemployed, causing their relative wage to fall "immensely."

A survey of 11,000 Americans found that most people expect the "substantial" scenario to play out, with GDP 10% higher and unemployment rising to around 5% by 2030. However, around 10% of respondents had views in line with the "extreme" scenario. The researchers emphasize that the distribution of AI's benefits and costs, such as retraining, income support, or universal basic income, will be crucial in shaping the economic landscape.