Press "Enter" to skip to content

MIT Index Shows AI Can Replace 11.7% of US Workers

MIT researchers have introduced the Iceberg Index, a new metric designed to measure how artificial intelligence technically overlaps with human job skills across the U.S. labor market.

Using a national-scale simulation of 151 million workers, the index finds that AI can already perform tasks representing 11.7% of wage value – about $1.2 trillion – largely in administrative, financial and professional services roles.

That exposure is five times greater than the visible AI disruption concentrated in technology occupations, which accounts for just 2.2% of wage value.

The study warns that traditional indicators such as GDP and unemployment explain less than 5% of the variation in AI-driven skills exposure, leaving policymakers blind to where disruption is forming.

By modeling how AI capabilities spread across regions and industries, Project Iceberg aims to give states a forward-looking tool to target reskilling and training investments before impacts fully materialize.

Read the MIT paper.

×