MIT模拟1.5亿打工人：程序员竟是安全？AI首波「血洗」1.2万亿白领薪资

Core Insights - A recent MIT study reveals that AI is poised to replace a significant portion of the U.S. workforce, particularly among white-collar workers, rather than just in the IT sector [1][9][18] - The study estimates that current AI technologies could replace 11.7% of the U.S. labor force, equating to a wage impact of approximately $1.2 trillion [1][9] - The research highlights that the majority of AI's impact will be felt in sectors such as finance, healthcare, human resources, logistics, and administrative roles, rather than in technology-focused jobs [9][10][18] Group 1: Research Methodology - MIT and Oak Ridge National Laboratory created a digital workplace simulator to analyze the U.S. labor market, utilizing one of the fastest supercomputers [3][6] - The simulation involved 151 million "agents," each representing a real U.S. worker, with detailed attributes such as location, job type, and skills [6][8] - The study employed the "Iceberg Index" to quantify the potential impact of AI on various job roles, revealing that only a small fraction of the impact is visible in high-profile tech layoffs [9][12] Group 2: Findings and Implications - The study found that only 2.2% of the total wage impact from AI (approximately $211 billion) is related to technology and IT jobs, indicating that the majority of job displacement will occur in less visible sectors [9][10] - The research serves as a "sandbox exercise," providing insights for governments and businesses to prepare for workforce transitions and identify which roles may need retraining [13][16] - The findings suggest that workers in routine information processing roles are at a higher risk of being replaced by AI, emphasizing the need for skill upgrades [18][19] Group 3: Technological Framework - The simulation's accuracy is supported by the AgentTorch platform, which allows for the modeling of complex interactions among millions of independent agents [19][21] - AgentTorch's differentiable modeling capabilities enable not only simulation but also optimization, helping policymakers identify effective solutions [22][23] - The platform's ability to connect macroeconomic data with individual skill sets allows for precise predictions of employment changes at a granular level [23][25]