The United States currently lacks the statistical infrastructure needed to accurately measure how many firms are using artificial intelligence, what they're using it for, and how many workers are deploying AI in their jobs — critical blind spots at a moment when policymakers are racing to understand AI's sweeping economic consequences. A new report from the Economic Innovation Group by Nathan Goldschlag argues that while the foundational statistical infrastructure already exists to answer these fundamental questions, it requires fast and significant upgrades to match the scale of the challenge. Without improved measurement capabilities, the report warns, policymakers risk getting their policy responses to AI wrong.
The report doesn't present new survey data or statistics about current AI adoption rates. Instead, it identifies a measurement gap in the nation's economic data collection systems. The report focuses on the need for U.S. statistical agencies to develop the capacity to track AI usage across firms and workers with accuracy, detail, and comprehensiveness. The analysis emphasizes that partial answers to questions about AI adoption and deployment won't be sufficient given the technology's rapid advancement and broad economic implications.
According to the report, researchers need accurate and comprehensive answers to several fundamental questions to properly track and understand AI's economic effects: how many firms are using artificial intelligence, what specific applications they're using it for, how many workers are using AI tools, and how employees are integrating AI into their work. The author argues that these aren't optional data points but essential measurements for understanding the technology's impact on workers and businesses. The report frames the challenge as urgent, noting that policymakers are already struggling to catch up with AI's consequences and can't afford to operate without reliable data.
The report's central argument rests on the premise that good policy depends on good data. Without robust measurement systems, policymakers operate blind — unable to see which industries are being transformed most rapidly, which workers face the greatest disruption, or which regions are benefiting from AI adoption versus being left behind. The report suggests that the existing U.S. statistical infrastructure, built through agencies like the Census Bureau and Bureau of Labor Statistics, provides a strong foundation but wasn't designed for the speed and scope of AI's economic transformation. The analysis implies that incremental improvements won't cut it; the measurement systems need substantial investment now to capture AI's effects in real time rather than years after the fact.
The report calls for necessary investments in U.S. statistical agencies to give them the ability to answer the most pressing questions about AI's impact on the American economy. The recommendation positions statistical infrastructure as both urgent and achievable — not a theoretical wish list but a practical upgrade to systems already in place. The bottom line is clear: America's policymakers can't manage what they can't measure, and right now they're flying blind on one of the most consequential economic shifts in generations.