MIT’s Iceberg Index: Mapping AI Exposure Across the U.S. Labor Market
Why have Technocrats always wanted to destroy labor? Howard Scott said in 1935, “technology has swept away the human worker” and then concluded that labor is non-essential to industry. Now, they have created a “digital twin for the U.S. labor market” only to watch it wither and die. Digital twinning is a AI-based tech to create simulated animations in silica that can be tweaked and experimented on, then changes are transmitted back to the physical twin. ⁃ Patrick Wood Editor.
Massachusetts Institute of Technology released a study showing artificial intelligence can already replace 11.7% of the U.S. labor market, equal to as much as $1.2 trillion in wages across finance, health care and professional services. The finding comes from a simulation platform called the Iceberg Index, developed with Oak Ridge National Laboratory. The index treats labor as a system to be modeled and probed rather than a static statistic.
The Iceberg Index models 151 million U.S. workers as individual agents with tagged skills, tasks, occupations and locations. It maps more than 32,000 skills across 923 occupations in 3,000 counties and measures which of those skills current AI systems can already perform. That level of granularity is what sets this effort apart from broad automation estimates.
Prasanna Balaprakash, ORNL director and co-leader of the research, described the project plainly: “Basically, we are creating a digital twin for the U.S. labor market,” and he pointed to ORNL’s Frontier supercomputer as the horsepower behind the simulations. The index runs population-level experiments to show how tasks, skills and labor flows could shift long before those changes appear in employment data. The goal is to offer a sandbox for testing policies before real dollars are spent.
The visible part of the disruption — layoffs and role shifts in tech and IT — represents only 2.2% of total wage exposure, or about $211 billion, according to the study. Beneath the surface lies the rest: routine and administrative functions across HR, logistics, finance and office administration that amount to the $1.2 trillion exposure. Those are the roles often missed by early automation forecasts.
The index is not a crystal ball predicting exactly when jobs will disappear, the researchers stress; it’s a skills-centered snapshot of what today’s AI systems can already do. Policymakers can use it to run what-if scenarios that explore training budgets, technology adoption rates and local economic impacts. That approach aims to reduce the risk of misallocating billions before testing interventions.
State partnerships helped validate the platform and turn it into a practical policy tool. Tennessee, North Carolina and Utah provided labor data and ran simulations to tune the model, and Tennessee cited the index in an official AI Workforce Action Plan released this month. Utah’s leaders are preparing a similar state-level report informed by the platform’s findings.
North Carolina state Sen. DeAndrea Salvador said the value is the local detail the model exposes, noting the ability to drill down to community-level skills and automation risk. “One of the things that you can go down to is county-specific data to essentially say, within a certain census block, here are the skills that is currently happening now and then matching those skills with what are the likelihood of them being automated or augmented, and what could that mean in terms of the shifts in the state’s GDP in that area, but also in employment,” she said. That granular visibility is what drew her to the project.
The simulations show exposure spread across all 50 states, including inland and rural regions that often get left out of the AI conversation. To help fill that gap, the Iceberg team built an interactive environment where states can toggle policy levers, shift workforce dollars and test different training mixes. The platform is designed so leaders can prioritize investments where they will do the most good.
The project report puts the capability bluntly: “Project Iceberg enables policymakers and business leaders to identify exposure hotspots, prioritize training and infrastructure investments, and test interventions before committing billions to implementation.” Balaprakash, who serves on Tennessee’s Artificial Intelligence Advisory Council, shared state-specific findings with the governor’s office and the state’s AI director. He highlighted that core sectors like health care, nuclear energy, manufacturing and transportation still rely heavily on physical work, offering some insulation from purely digital automation.
For now, the team positions Iceberg as a sandbox rather than a finished product: “It is really aimed towards getting in and starting to try out different scenarios,” Salvador said. The platform gives states a chance to experiment with policy choices and training strategies before they scale up expensive programs. That testing-ground approach is the selling point for officials wrestling with the scale and speed of AI change.