Generative AI’s Role in Preventing US Recession

Defying predictions: The three forces behind US economic strength in 2025

In this commentary, Pinelopi Koujianou Goldberg argues that the US economy has proven more resilient than economists predicted in 2025 due to a combination of front-loading behavior by businesses and consumers anticipating tariffs, substantial investment in AI infrastructure providing a structural growth boost, and the eventual moderation of tariff policies that reduced uncertainty.

High hopes for AI have fueled most of the stock-market gains since last year. While many fear that the bubble could burst at any moment, anyone who has used large language models over the past two years will have seen not only what they can already do, but also the immense and rapid progress they have made.

However, this strength may be temporary rather than reflecting fundamental economic robustness, as the lagged effects of trade restrictions and policy uncertainty are beginning to manifest in slowing investment, weakening labor markets, and shifting consumption patterns.
Why the US Economy Has Remained Stronger Than Expected
By: Pinelopi Koujianou Goldberg – Yale University

Generative AI is reshaping US productivity: new evidence from Worker Surveys

According to Alexander Bick, Adam Blandin, and David Deming, Generative AI adoption among US workers has surged from 44.6% in August 2024 to 54.6% in August 2025, representing a faster adoption rate than personal computers or the internet experienced at comparable stages, with particularly strong growth in nonwork settings (36% to 48.7%). Workers report that generative AI saves them approximately 1.6% of all work hours, which translates to potential aggregate labor productivity gains of up to 1.3% since ChatGPT’s release—findings consistent with observed productivity growth exceeding pre-pandemic trends by 1.89 percentage points. Quarterly date from the Real-Time Population Survey show:

• Work vs. nonwork usage: While work adoption increased from 33.3% to 37.4%, nonwork adoption grew even more rapidly from 36% to 48.7%, indicating broader consumer engagement beyond professional settings.
• Intensity of use: Workers now spend 5.7% of their work hours using generative AI (up from 4.1% in November 2024), with users reporting time savings equivalent to 1.6% of all work hours when nonusers are included.
• Industry-level correlation: Industries with higher AI adoption and time savings show stronger productivity growth relative to pre-pandemic trends, with a 1 percentage point increase in time savings associated with 2.7 percentage points higher productivity growth, though causality cannot be definitively established.
• Productivity implications: The technology may have already increased labor productivity by up to 1.3%, contributing to the 2.16% annualized productivity growth from Q4 2022 through Q2 2025—significantly above the 1.43% pre-pandemic average.

The State of Generative AI Adoption in 2025
By: Alexander Bick, Adam Blandin, David Deming – Federal Reserve Bank of St Louis
(The above article is an updated version of the “The Rapid Adoption of Generative AI” paper highlighted in our newsletter “How fast is generative AI being adopted?” )