Total factor productivity and generative AI
Amidst all concerns about the potential dangers associated with large language models such as ChatGPT, Baily, Brynjolfsson, and Korinek, in this commentary, strike an optimistic note on the potential of generative AI to enhance total factor productivity (TPF). They make some interesting points:
- Contrary to common beliefs, digital technologies did indeed lead to a surge in TPF during the period between 1995-2005 when a huge wave of investment in computers and communications transformed businesses. It was slow growth together with the outsized profits of tech companies, that resulted in skepticism about the benefits of digital technologies for the broad economy.
- Generative AI has the potential to produce similar increases in TFP on specific occupations or tasks: for example, recent research shows that software engineers can code up to twice as fast using a tool called Codex, based on the previous version of the large language model GPT-3; writing tasks can also be completed twice as fast enabling economists to be 10-20% more productive using large language models.
- There are two channels through which AI can boost productivity: The first channel is the increased efficiency of output production. The second, and most important, is the acceleration of innovation and thus future productivity growth. Cognitive workers produce current output but also generate the technological progress that boosts future productivity. “If productivity growth was 2% and the cognitive labor that underpins productivity growth is 20% more productive, this would raise the growth rate of productivity by 20% to 2.4%. In a given year, such a change is barely noticeable and is usually swamped by cyclical fluctuations.”
- Usually new technologies deliver productivity gains only after a period of investment in complementary intangible goods, such as business processes and new skills. But cognitive automation can often be rolled out quickly via software: ChatGPT was the most rapid product launch in history—it gained 100 million users in just two months—because it was accessible to anyone with an internet connection and did not require any hardware investment on the users’ side.
Machines of mind: The case for an AI-powered productivity boom
Authors: Martin Neil Baily, Erik Brynjolfsson, Anton Korinek
From: Center on Regulation and Markets, Stanford University
What do economists in the U.S. and Europe think?
But Baily, Brynjolfsson, and Korinek’s optimism is not broadly shared: surveys run by the Chicago Booth Clark Center for Global Markets show that economists both in the U.S. and Europe are less sanguine about the positive impact of AI on productivity.
U.S. economists:
European economists:
AI and Productivity Growth – US
AI and Productivity Growth – Europe
From: Chicago Booth – Kent A. Clark Center for Global Markets
