In late March 2023, just months after ChatGPT’s public launch, Goldman Sachs economists Joseph Briggs and Devesh Kodnani published a report titled “The Potentially Large Effects of Artificial Intelligence on Economic Growth.” Summarized on the firm’s own site under the headline “Generative AI could raise global GDP by 7 percent,” it became one of the most-quoted economic forecasts of the generative-AI era.
The report’s most-repeated figure is that shifts in workflows triggered by generative AI could expose the equivalent of 300 million full-time jobs to automation worldwide. The authors estimated that roughly two-thirds of US occupations are exposed to some degree of AI automation, and that of the exposed work, between a quarter and as much as half of the workload could be replaced. On the upside, they projected that widespread adoption could eventually raise annual global GDP by about 7 percent, roughly 7 trillion dollars over a ten-year horizon, and lift labor productivity growth by about 1.5 percentage points per year.
The “300 million jobs” line entered the public conversation almost immediately and was frequently misreported as 300 million jobs lost rather than exposed, a meaningful distinction since exposure includes work that is augmented rather than eliminated. Skeptics, including Daron Acemoglu, later cited these very numbers as examples of forecasts they considered too bullish. The report is a landmark less for being correct than for crystallizing the scale of the economic stakes in the months after ChatGPT arrived.