A new MIT study indicates artificial intelligence can perform work equivalent to 11.7% of U.S. jobs, potentially affecting 151 million workers.
The research, conducted by MIT and Oak Ridge National Laboratory under Project Iceberg, estimates current AI systems could manage tasks representing 11.7% of the U.S. labor market. This figure equates to approximately $1.2 trillion in wages. This analysis differs from previous estimates by focusing on jobs where AI performs tasks at a cost competitive with or cheaper than human labor.
Project Iceberg developed a labor simulation, described as a “digital twin of the U.S. labor market,” according to Prasanna Balaprakash, a study co-leader and director at Oak Ridge National Laboratory. This model simulates 151 million workers with specific skills, occupations, and locations. It tracks over 32,000 skills across 923 job types in 3,000 counties, mapping them against current AI capabilities.
The 11.7% figure reflects AI’s technical capability and economic feasibility, not a prediction of immediate job displacement. AI adoption has largely focused on tech work, particularly coding, representing about 2.2% of wage value, or approximately $211 billion in pay. Researchers found AI is capable of handling cognitive and administrative tasks in finance, healthcare, and professional services, collectively representing around $1.2 trillion in wages. This is about five times the currently observed impact.
Significant exposure is noted in white-collar, knowledge-intensive fields. These include finance, healthcare administration, human resources, logistics, and professional services such as legal and accounting. Existing AI tools, including large language models (LLMs), can execute many routine tasks in these areas. Much of this potential disruption is concentrated in traditional back-office and professional roles.
MIT researchers caution that capability does not automatically translate to widespread job losses. Earlier work from MIT’s Computer Science and Artificial Intelligence Laboratory indicated that fully replacing human workers with AI remained too expensive or impractical in the near term for many roles. Separate research from MIT Sloan concluded that AI exposure from 2010 to 2023 did not lead to broad net job losses and often coincided with faster revenue and employment growth at adopting firms.
The Iceberg Index serves as a tool for policymakers and business leaders to evaluate scenarios before committing resources. Tennessee, North Carolina, and Utah are already using the platform to assess AI’s potential impact on their workforces and to inform state-level AI workforce action plans, as stated in the MIT report.
The study highlights a narrowing timeframe for companies to address AI as a future issue. For governments, it raises questions about worker retraining, supporting affected regions and sectors, and adapting tax and social safety net systems for a labor market where software can perform a meaningful share of work.
