MIT Study: AI Tools Now Cover Skills Worth $1.2 Trillion of the U.S. Wage Bill

A major new study from MIT quantifies what many have suspected: AI tools are not on the horizon — they are already embedded in the fabric of American work.

The paper, titled The Iceberg Index: Measuring Skills-centered Exposure in the AI Economy, was led by MIT researcher Ayush Chopra alongside nine co-authors from MIT, Oak Ridge National Laboratory, Project Iceberg, and state policy bodies in North Carolina and Utah. The team mapped 32,000 skills across 923 occupations representing 151 million workers in roughly 3,000 U.S. counties, using O*NET and American Community Survey data.

On the AI side, researchers catalogued more than 13,000 tools — including workplace copilots, automation platforms, and specialized systems — and matched their capabilities to human occupational skills. The result: current AI tools already overlap with skills that represent 11.7% of the U.S. wage bill, equivalent to approximately $1.2 trillion per year.

The researchers call this an upper bound. Actual near-term exposure is likely lower when tasks require specialized knowledge, internal systems, or local judgment. The study measures skills exposure, not job elimination — but the scale of the finding matters for anyone planning AI strategy or workforce investment.

Key Takeaways:

• Current AI tools overlap with $1.2 trillion worth of U.S. annual labor output — 11.7% of the total wage bill
• The study mapped 13,000+ AI tools against 32,000 skills across 923 occupations and 151 million workers
• The 11.7% figure is an upper bound; real exposure depends on context, judgment, and specialized knowledge

Read the full article on Space Daily