Anthropic launches interactive tool to test AI's impact on U.S. economy

In one scenario, AI gives the economy a gentle boost, lifting growth rates with little effect on workers. In the other extreme, GDP soars while unemployment rises. Nearly 14% of workers lose their jobs to AI in that scenario, and less than half of them find new ones.

The outcomes depend on how capable and flexible the technology is, how quickly it is adopted, whether it supports or replaces workers, and whether displaced workers find new work. Anthropic experts write in an accompanying blog post that in the modest change scenario, AI is a small technology and the economy continues on a normal path with AI making changes around the margins. In the extreme scenario, they write, AI transforms the economy, and the macroeconomic consequences go beyond any event in history in both magnitude and speed.

Authors of the Anthropic study do not offer a prediction of which outcome is more likely. They say AI experts generally project a more rapid spread of the technology while economists tend to be more cautious. Anthropic co-founder Jack Clark is somewhere in the middle. "I think the technology will keep developing at a very, very fast and sustained rate but diffusion of the technology will likely be more challenging than people think," Clark told NPR. "So it will get really, really good. But it will make its way into the economy more slowly."

A survey by Anthropic of nearly 11,000 people found that public expectations of AI are also mixed. The average respondent projects a significant boost in productivity and economic growth but also substantial disruption to workers in AI-sensitive fields.

Anton Korinek, Anthropic's head of transformative AI economic studies, says a major determinant of AI's impact will be how quickly it spreads through the economy. "If the AI can do amazing things but nobody uses it, then it's not going to have an economic impact," Korinek says.

While rapid adoption of the technology could be more disruptive, it could also produce much faster economic growth. In Anthropic's extreme case, GDP grows at more than seven times its current pace. That could produce a lot of additional tax revenue to support workers who are hurt by AI. Clark says that if the economy ends up with this level of GDP growth, policymakers would have moves available that are unimaginable today. "Policymakers should get ready to spend," he says.