A MIT study, conducted with Oak Ridge National Laboratory and reported by CNBC, reveals that artificial intelligence can already replace 11.7% of the U.S. workforce, representing $1.2 trillion in wages. The analysis is based on the Iceberg Index, a simulation tool that acts as a digital twin of the labor market: it models 151 million American workers as individual agents, each with their own skills, tasks, occupation, and location, and maps 32,000 skills across 923 occupations in 3,000 counties.
The iceberg metaphor structures the findings. The visible portion of the impact (2.2% of the workforce, $211 billion) corresponds to layoffs already observed in tech and IT. The hidden portion, far more massive (9.5%), concerns routine functions in human resources, logistics, finance, and administration — jobs that are exposed but whose disruption does not yet appear in the statistics. Contrary to common assumptions, the impact is not limited to coastal tech hubs: all 50 states are affected, including rural regions often overlooked in discussions about AI.
The large-scale computation is made possible by the Frontier supercomputer at Oak Ridge National Laboratory. The tool stands out from traditional approaches through its granularity (county-level, even zip-code-level analysis), its proactivity (identifying disruptions before they appear in economic data), and its experimentation capability: policymakers can test "what-if" policy scenarios before implementation.
Several states are already adopting it. Tennessee has integrated the Iceberg Index into its AI Workforce Action Plan, targeting healthcare, nuclear energy, manufacturing, and transportation, with a strategy of using robotics and AI assistants to strengthen rather than hollow out industries. Utah is preparing a report based on the tool to identify exposure hotspots and prioritize investments. North Carolina is working closely with MIT researchers on a county-level analysis of skills at risk of automation and the potential impact on the state's GDP.
The researchers position the Iceberg Index as an experimentation "sandbox" for continuous improvement, not as a finished product. The study nonetheless marks a turning point: it shifts preparation for AI-driven labor market transformation from a reactive approach to a proactive, granular, and data-driven strategy.