A new MIT report using Project Iceberg, a labor‑market 'digital twin' built with Oak Ridge National Laboratory, estimates current AI systems can economically perform tasks equal to about 11.7% of U.S. wage value — roughly $1.2 trillion. Researchers including Prasanna Balaprakash say the model maps 32,000 skills across occupations and finds AI already competitive on many cognitive and administrative tasks in finance, healthcare and professional services. The paper stresses the figure reflects technical and economic feasibility, not an immediate job‑loss timetable, and calls on policymakers and educational institutions to accelerate reskilling, curriculum updates and workforce transitions.