Iceberg Index: AI can already replace 11.7 percent of US workers

Contents

11.7 percent of US workers could already be replaced by artificial intelligence today. This is the conclusion of a recent study from the Massachusetts Institute of Technology (MIT). MIT researchers, together with the Oak Ridge National Laboratory, have developed a simulation tool called "Iceberg Index" to create a digital twin of the US labor market. This allows them to research the effects much more broadly than has been practiced in previous surveys, they say.

As MIT shows in the study, this concerns a payroll of around 1.2 trillion US dollars that would be lost due to the use of AI. The financial sector, healthcare, and corporate services are particularly affected.

Digital Twin of the Labor Market

The Iceberg Index simulates the interactions of 151 million US employees and analyzes which of their tasks current AI technologies can take over. The tool captures over 32,000 individual skills across 923 job profiles. It allows the effects to be broken down to individual regions so that, for example, the impact on individual cities can be determined.

The simulation runs on the Department of Energy's Frontier supercomputer in Tennessee and treats each employee as an individual agent with specific skills, tasks, job title, and location, as US television broadcaster CNBC reports. The simulation tool is intended, for example, to enable decision-makers to foresee developments in AI automation before actual waves of layoffs begin. This could lead to political decisions being made to mitigate the effects. Furthermore, it is possible to test the use of further training programs or retraining in the simulation.

Tech Industry Only the Tip of the Iceberg

The results differ significantly from public perception: The visible waves of layoffs in the tech industry – for example, at Amazon, Salesforce, and others – represent only 2.2 percent of economic power from wage payments, about 211 billion US dollars. For society, however, it is much more threatening when routine tasks in human resources, logistics, financial administration, and office organization are converted to AI. These areas are often overlooked in other automation forecasts but affect larger segments of the population. Potentially impacted professions are found in all 50 states, including rural inland regions.

Comparison to German Forecasts

The MIT results complement the German research of the Institute for Employment Research (IAB), which also forecasts significant upheavals. While the IAB speaks of 1.6 million jobs that will be lost or created in the next 15 years, the Iceberg Index focuses on automation potentials that can already be measured today. The difference lies in the methodology: the MIT tool does not simulate future scenarios but depicts what current AI systems can technically already achieve.

(mki)