AI & Economic Measurement

Artificial intelligence (AI), combined with increasingly rich datasets, is reshaping how economists measure key aspects of the economy. AI methods are also changing how we collect data, construct statistics, and evaluate policy. To explore these issues, the NBER will host a conference on May 7, 2026, that will gather researchers studying the methodology and application of AI for economic measurement. The conference will be organized by Erik Brynjolfsson (Stanford) and Karen Dynan (Harvard).
The conference will bring together leading academics who are working on AI, researchers from the frontier labs in industry, and experts on economic measurement who are very familiar with national income accounts and price measurement. The conference will be organized around three themes: i) how AI tools can improve the construction of traditional economic statistics, ii) economic measurement of the effects of AI in the economy, including potential modifications to measures of labor market activity or productivity that might be appropriate to capture the effects of AI; and iii) how new information generated by AI tools may provide information on economic activity more broadly. The development of new AI-related indicators, such as the index developed in 2025 by Anthropic, would fall within the third theme.