Measuring the AI Economy Before GDP Can See It: A Review and Extension of Korinek and McKelvey

This release contains a standalone review-and-extension note on Anton Korinek and Patrick McKelvey’s “Measuring the AI Economy.” The article treats the paper as an important early contribution to AI measurement infrastructure, focusing on its AI GDP boundary, production-side compute methodology, and quality-adjusted output estimates. It argues that the headline 2,000–2,600 percent growth figures should be read as AI productive-capacity signals rather than ordinary GDP or welfare growth. The note extends the discussion by formalizing the Capacity–Realization Gap and proposing sensitivity-banded AI satellite accounts.

Keywords: artificial intelligence; AI GDP; national accounts; satellite accounts; productivity measurement; compute; model capital; AI economics; uncertainty bands; capacity-realization gap; quality-adjusted output; AI measurement infrastructure