May 30, 2024 Amid growing excitement about artificial intelligence’s potential to revolutionize productivity and economic growth, an MIT professor offers a sobering perspective. Daron Acemoglu, a respected economist, presented his insights in a paper for the US National Bureau of Economic Research, cautioning that AI might actually exacerbate income inequality rather than turbocharge economic output.
In his paper, “The Simple Macroeconomics of AI,” Acemoglu critiques the optimistic forecasts that AI will lead to significant GDP growth and a surge in blue-collar wages. While some projections, including a notable one from Goldman Sachs, suggest AI could add nearly $7 trillion to the global GDP over the next decade, Acemoglu remains skeptical. He argues that such forecasts often overlook the nuanced impacts on the macroeconomy, particularly concerning wages and the income gap between capital owners and workers.
Drawing from research estimating that AI could automate about 20% of tasks currently performed by US workers, Acemoglu calculates potential labor cost savings. However, he contends these savings might not translate into broad productivity gains. He predicts only a modest increase in total factor productivity (TFP) growth, potentially leading to GDP growth of just 0.93% to 1.16% over the next decade—far less than the more optimistic forecasts.
Moreover, Acemoglu warns that the economic benefits of AI could be offset by significant costs, such as increased energy consumption. He also raises concerns that AI-driven automation might not lead to wage improvements for low-end and middle-performing workers, potentially deepening existing inequalities.
This critical examination by Acemoglu invites policymakers and business leaders to temper their expectations regarding AI’s economic impact. It also underscores the need for careful consideration of how AI technologies are deployed to ensure they contribute positively to societal welfare without widening the economic divide.
