Developed jointly by MIT and Oak Ridge National Laboratory, the "Iceberg Index" labor simulation tool creates digital twin models for 151 million U.S. workers. According to research lead Prasanna Balaprakash, this innovative tool systematically quantifies for the first time the real-world effects of AI technology and related policies on the job market by simulating the interactive dynamics of the entire U.S. labor force. As a breakthrough study, it integrates analysis of over 32,000 skills across 923 occupations and evaluates the current practical capabilities of AI systems in various sectors.\n\nDespite the technology sector continuing to attract capital and public attention, the MIT study reveals that AI’s actual penetration is far broader. Everyday functions in human resources, logistics, finance, and office administration have become key targets for AI substitution. These areas represent a combined wage pool of approximately $1.2 trillion — a significant portion of the U.S. labor market that traditional automation models have often overlooked. This finding highlights a clear industry trend: AI is moving from high-tech fields into routine office work.\n\nResearch results based on the Iceberg Index are already influencing labor policy development across U.S. states. Tennessee has cited the study’s data in its official AI workforce action plan, and Utah plans to release a related analysis report.

The research team has established partnerships with multiple state governments, using forward-looking simulations to help policymakers anticipate the differentiated impacts of AI on local job markets, providing a scientific basis for resource allocation and policy adjustments.\n\nThe study challenges the common perception that "AI risk is concentrated in coastal tech hubs," showing that AI-affected occupations exist in all 50 states — with even more pronounced effects in inland and rural areas. This finding offers a fresh perspective on the regional impact of AI and reveals new forms the digital divide may take in the AI era. As large model technology becomes more widespread, the growing demand for computing power is driving AI infrastructure to expand into more regions, further reshaping the employment distribution landscape.\n\nTo address the labor market transformation brought by AI, the research team has built an interactive simulation environment that allows state decision-makers to test the effects of different policies on local employment and the economy before committing real resources. This data-driven policy experimentation approach represents an important innovation in governance models for the AI era and provides a replicable framework for other countries. As chip technology continues to break through, the sustained improvement of AI computing power will further accelerate this process, presenting increasingly complex challenges for policymaking.