Microsoft Shares Daron Acemoglu's Cautious AI Forecast

Daron Acemoglu, Nobel Prize-winning economist, describes a measured trajectory for AI: in his view, total automation is less convincing than tools that extend human skills. He emphasizes unresolved organizational constraints and estimates the expected impact on GDP and employment at modest levels. Microsoft shares this position on its blog, The Humanist Review of AI.

Acemoglu Favors Augmented AI Over Total Automation

Daron Acemoglu argues that artificial intelligence systems designed to extend human skills could offer more productivity gains than complete automation. He points out that a 99% accuracy rate is often insufficient due to unforeseen circumstances and the actual needs of users. According to him, what is primarily lacking are easy-to-deploy applications that would concretely change production methods.

The Barrier Comes from Organizations, Not Model Size

For Daron Acemoglu, the main obstacle to productivity related to artificial intelligence is human: tasks need to be redistributed, employees need training, and companies must be restructured before any gains can be seen. He estimates that this adaptation could take longer than electrification. He also believes that increasing the size of models will not resolve this blockage as long as organizational adjustment is not achieved.

Moderate Projections Published by Microsoft

Daron Acemoglu predicts that artificial intelligence will only increase GDP by about 1.5% over a decade and will replace at most 5% of jobs, levels that are significantly lower than some projections from AI labs. He also acknowledges that the pace of AI development remains difficult to anticipate. Microsoft published this analysis on its corporate blog, The Humanist Review of AI, where articles are signed by hand.