The Measurement Revolution? Credible Measurement and Inference in the Age of AI
2026-08-24 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors explain how artificial intelligence (AI) is changing the way economists measure things by turning messy data like text and pictures into clear numbers more cheaply and on a big scale. This means researchers now have many ways to measure the same thing and must carefully decide which is best. They outline three stages where AI is used and stress that to trust AI-based measurements, researchers need strong validation by comparing against clear standards. The authors also discuss how even biased AI predictions can still be useful with the right validation, and what to do if random validation data isn't available.
artificial intelligencemeasurementeconomicsvalidationstructured dataunstructured databiasinferenceproxy variablesempirical research
Authors
Melissa Dell, Ashesh Rambachan
Abstract
Artificial intelligence (AI) is transforming measurement in economics. AI models convert unstructured data, such as text and images, into structured variables at low cost, making previously prohibitive measurement feasible at scale. This shifts the bottleneck from finding any scalable measure of a phenomenon to choosing among many plausible ones, which may support different empirical conclusions. This review provides guidance for navigating that shift. We describe three stages at which AI enters the measurement pipeline---discovery, construct definition, and observation---and what each demands of researchers. We argue that credible inference with AI-generated variables requires appropriately designed validation: anchoring measurement to explicit criteria, rather than informal claims that a proxy is reasonable. We then examine how validation samples support valid inference even when AI predictions are arbitrarily biased, and what can be done when a random validation sample is unavailable.