The Federal Reserve Board published a research paper proposing a new firm-level measure of artificial intelligence research and development, the AIR Index, based on semantic similarity between earnings conference call transcripts and leading AI research papers. The paper links higher AIR readings to near-term market valuation effects and subsequent increases in capital expenditure, while finding no significant effects on productivity or employment. The AIR Index varies widely across industries, with sustained strength in computer and electronic manufacturing and faster growth in computing infrastructure and educational services after the introduction of ChatGPT in November 2022. Higher AIR scores are associated with an immediate increase in Tobin's Q and help explain cross-sectional differences in cumulative absolute returns following conference calls, suggesting investors respond to substantive AI discussions. A sharp rise in the AIR Index is followed by persistently higher year-over-year capital expenditure growth for about a year before tapering, and industry-level survey data from Census show recent AIR growth correlates with broader AI adoption trends.
Federal Reserve Board 2025-02-24
Federal Reserve Board research introduces an AIR Index to measure firm-level AI R&D from earnings call transcripts
The Federal Reserve Board released a research paper introducing the AIR Index, a firm-level measure of AI research and development, correlating with market valuation effects and capital expenditure increases. The index shows significant variation across industries, with notable strength in computer and electronic manufacturing and growth in computing infrastructure and educational services post-ChatGPT. Higher AIR scores are linked to immediate increases in Tobin's Q and explain differences in cumulative returns, indicating investor sensitivity to AI discussions.