Economists Warn AI Impact Metrics Are Unreliable

Over 200 economists, including 16 Nobel laureates and the chief economists of OpenAI and Anthropic, signed a statement on Monday warning that the profession lacks the tools to measure AI's economic impact. The declaration, "We Must Act Now," notes AI could trigger an economic transformation larger than the Industrial Revolution over the next decade, but current data infrastructure is dangerously inadequate.

For investors and corporate strategists trying to price AI risk, the admission is stark. Anton Korinek, a University of Virginia professor and organizer, said: "We are driving in the fog, and it is extraordinarily difficult to anticipate what will happen next." Nela Richardson, ADP’s chief economist, dismissed much of the current public debate over AI employment effects as "guesswork."

The confusion stems directly from the underlying measurements. Apollo Global Management chief economist Torsten Slok highlighted five competing frameworks used to define "AI exposure," each yielding drastically different results. The methods range from analyzing real-world Claude and Microsoft Copilot usage logs to asking ChatGPT to grade its own usefulness. "When someone says a job is 'highly exposed to AI,' the honest first question is: Exposed by which measure, and measuring what?" Slok wrote. "Until that is pinned down, the label 'AI exposure' carries far less meaning than it appears to."

Theoretical frameworks consistently overstate risk compared to actual usage data, particularly for high-profile roles like writers, tax preparers, and telemarketers. However, relying solely on aggregate headline labor figures carries its own risks, as macro data can mask localized disruptions. Stanford’s Erik Brynjolfsson pointed to his Canaries Dashboard, built with ADP Research, which tracks 4.6 million workers across more than 730 occupations. It shows employment for workers aged 22 to 25 in AI-exposed occupations shrinking more than 4% annually, even as aggregate labor figures remain calm.

Michael Spence, a Nobel laureate at NYU, called for an "all hands on deck" approach to steering AI. Daron Acemoglu, an MIT Nobel laureate and noted AI productivity skeptic, signed the statement to urge a redirection of the technology, saying he was "so happy to join other leading experts in calling for the urgent need to redirect AI so that its risks are minimized and it can work for the benefit of workers and society."

"AI capabilities are advancing far faster than our understanding of the economic implications," Brynjolfsson said. "We are flying blind into one of the most consequential periods in world history. We need timely, trusted evidence to understand where AI is creating value and where it is disrupting work."