The true cost of using AI is not the price of generating a document but the time it takes for a human to check if it is wrong.
Productivity gains from AI are limited by the high cost of verification rather than the low cost of creation. High-value tasks remain expensive because they require expert human oversight to ensure accuracy and safety. Many people expect AI to make all white-collar work significantly cheaper and faster almost overnight. This microeconomic analysis proves that the bottleneck has simply shifted from the doing of work to the checking of work. AI will only be truly transformative for tasks where the output can be automatically and cheaply verified.
Verified Output and the Microeconomics of AI Adoption Why AI Productivity Depends on Verification, Not Generation Alone
SSRN · 6701098
Artificial intelligence sharply reduces the cost of generating cognitive output. It can produce text, code, classifications, summaries, analyses, recommendations, and decision support at a speed and scale that previously required scarce human labor. Yet generated output is not the same as economic output. In consequential settings, firms can use AI output only after it has been verified, authorized, and made reliable enough for use under liability. This paper develops a microeconomic framework f