MIT research finds: Artificial intelligence can already replace 11.7% of the U.S. workforce

```

A study released by MIT on Wednesday shows that artificial intelligence is already able to replace 11.7% of the U.S. labor market, which is equivalent to as much as $1.2 trillion in wages in the finance, healthcare, and professional services sectors.

This study used a workforce simulation tool called the “Iceberg Index,” jointly developed by MIT and Oak Ridge National Laboratory (ORNL). ORNL is a Department of Energy research institution located in eastern Tennessee, home to the Frontier supercomputer used for large-scale modeling.

The Iceberg Index can simulate interactions among 151 million U.S. workers and assess how they might be affected by AI and related policies. The version of the Iceberg Index released earlier this year offers a forward-looking perspective, showing how AI will reshape the U.S. labor market—not just in coastal tech hubs, but in every state across the country.

For U.S. lawmakers planning to invest billions in retraining and upskilling, the Iceberg Index provides a detailed map pinpointing regions facing potential structural shocks, with granularity down to the ZIP code.

Prasanna Balaprakash, co-leader of the research and ORNL director, said:

Essentially, we are creating a digital twin system for the U.S. labor market. This index can run population-level experiments to reveal how AI will transform tasks, skills, and workforce flows before any actual employment changes occur.

Balaprakash, a member of Tennessee’s AI Advisory Council, also shared state-level results with the governor’s team and state AI director.

He noted that Tennessee’s core industries—healthcare, nuclear energy, manufacturing, and transportation—still heavily rely on physical labor, making them relatively less susceptible to purely digital automation. The key question is how to use new technologies like robots and AI assistants to enhance rather than diminish these industries.

The Iceberg Index treats 151 million workers as independent agents, each with their own skills, tasks, occupations, and geographic information. It maps over 32,000 skills to 923 occupations and 3,000 counties, then assesses which skills current AI systems can perform.

Researchers found that the visible “tip of the iceberg”—such as layoffs and job changes in technology, computing, and IT industries—accounts for only 2.2% of total wage exposure (approximately $211 billion). The real, hidden exposure amounts to as much as $1.2 trillion, encompassing routine functions like HR, logistics, finance, and office administration, sectors often overlooked by traditional automation forecasts.

Researchers emphasize the Iceberg Index is not a tool for predicting specific “when and where” unemployment, but rather aims to present a “skills-based snapshot” showing what today’s AI systems can already do. This helps policymakers structurally explore various “what-if scenarios” before investing money or enacting legislation.

The research team is working with U.S. state governments to conduct early simulations. Tennessee, North Carolina, and Utah have used their state-level labor data to validate the model and begun designing policy scenarios using the platform. Tennessee took the lead by citing the Iceberg Index in its official “AI Workforce Action Plan” released this month. Utah is preparing a similar Iceberg-based modeling report.

DeAndrea Salvador, a North Carolina senator working closely with MIT, said one appeal of the research is its ability to reveal impacts that traditional tools can’t capture. One of the most valuable features is the ability to drill down to local-level data:

You can see data specific to counties or even census tracts, understand what skills are current, match those skills with their probability of being automated or augmented, and assess impacts on local GDP and employment changes. This kind of simulation is especially important as more states establish specialized AI task forces and research groups.

The Iceberg Index also challenges a common assumption—that AI risks will mainly concentrate in tech jobs at coastal centers. The simulations reveal AI-exposed occupations are spread across all 50 states, including inland and rural areas often left out of AI discussions.

To address this gap, the Iceberg team has built an interactive simulation environment where states can test different policy levers—from adjusting workforce budgets and optimizing training programs to exploring how changes in technology adoption will affect local employment and GDP.

The report states: “The Iceberg program helps policymakers and business leaders identify risk hotspots, prioritize training and infrastructure investment, and test interventions before investing billions in implementation.”

Currently, the team positions the Iceberg Index as a “sandbox” for states to prepare for AI’s impact on the workforce. The goal is to let everyone get started and try out different scenario plans.

Risk Warning and DisclaimerThe market has risks, investment needs caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their particular circumstances. Investment based on this is at your own risk. ```