AI can replace 11.7% of US labor market, study finds

A new study by the Massachusetts Institute of Technology found that AI can already replace 11.7% of the US workforce, equal to about US$1.2 trillion in wages across sectors such as finance, healthcare, and professional services.

Researchers developed a labor simulation tool, the Iceberg Index, with Oak Ridge National Laboratory to map how 151 million US workers interact with AI and related policies.

The tool analyzes over 32,000 skills across 923 occupations, highlighting that routine jobs in fields like human resources and logistics are often overlooked in automation forecasts.

Visible tech layoffs and role shifts account for just 2.2% of wage exposure, according to the study.

The Iceberg Index does not predict when or where jobs will be lost but provides a snapshot of what AI can already do and helps policymakers run simulations before making training or reskilling investments.

States including Tennessee, North Carolina, and Utah are using the platform to inform workforce policy decisions.

🔗 Source: CNBC

🧠 Food for thought

Implications, context, and why it matters.

The automation potential isn’t evenly distributed across occupations

Iceberg Index measures AI’s ability to automate tasks across 32,000 skills and 923 occupations, not only headline roles 1.
Routine roles in human resources or logistics face higher exposure, yet most press focuses on tech layoffs, which account for only 2.2% of wage exposure 1.
It covers what AI can do today without guessing when or where jobs may change, which helps separate capability from rollout timing 1.
Occupation-level mapping surfaces pockets of automation risk that high-level stats miss 1.

Public-sector AI workforce programs are creating near-term vendor opportunities

Tennessee, North Carolina, and Utah use the Iceberg platform to guide workforce policy choices 1.
New York’s AI Prep program will award contracts to up to two program administrators to build curriculum and run two tracks for underrepresented groups, with submissions due October 1, 2025 2.
The Advanced Robotics for Manufacturing (ARM) Institute seeks proposals for an automated system tied to RoboticsCareer.org to maintain a U.S. robotics and AI training database with almost 17,000 programs from over 2,400 providers 3.
Federal Workforce Innovation and Opportunity Act (WIOA) funding now covers AI skills training 4. That widens options for providers.