More than 200 economists and AI researchers, including 16 Nobel laureates, recently signed an open letter warning of “an unprecedented transformation of our economy, larger than the Industrial Revolution.” Organized by Stanford’s Digital Economy Lab, the letter implores policymakers to “act now” against mass job displacement. It joins a long line of AI predictions that have thus far, proven to be incorrect.
In May 2025, Anthropic CEO Dario Amodei told Axios that AI would eliminate half of all entry-level white-collar jobs and push unemployment to 10 to 20 percent within one to five years. Fourteen months later, data from the Bureau of Labor Statistics reveals that the unemployment rate in the United States remains stable at 4.2 percent. That metric certainly doesn’t look like a labor market that is cracking under automation.
The hiring data tells the same story. Ramp studied AI spending across more than 21,000 firms and found that the heaviest AI adopters expanded entry-level roles by 12 percent over the following two years. Contrary to popular belief, attempts to utilize AI can actually cost more than human labor. Last month, an NVIDIA executive noted that the cost of running AI is “far beyond” the cost of human workers.
While the long term effects of AI on the labor market remain to be seen, what the discrepancy between AI doomsday predictions and the current employment data reveals is that predicting the effects of a general-purpose technology like AI is nearly impossible. That is exactly why lawmakers shouldn’t legislate as if they have a crystal ball.
Economists struggle to predict anything beyond second-order effects. The first-order effect of rent control is lower prices. The second-order effect is that landlords, unable to charge market rents, become more selective about tenants and withdraw units from the market, leading to lower supply in the long run. Forecasting anything beyond that is notoriously unreliable, which is why macroeconomic forecasting is nearly impossible. A disruptive technology like AI compounds this uncertainty in the same way that chaos theory (or the butterfly effect) illustrates how small changes in initial conditions can produce unpredictable outcomes in the physical world.
AI could substitute for labor, but it could just as easily increase demand for it. Economists are increasingly comparing AI’s trajectory to Jevons’ paradox. The paradox is the observation that as a resource like energy becomes more efficient to use, people often end up using more of it, not less. Instead of becoming redundant, workers made more productive by AI may actually end up becoming more valuable to organizations.
Utah has embraced the potential of AI to transform our society, which is why it has created an AI sandbox to give companies the chance to demonstrate the value of their technology without unnecessary regulation. Currently, the Beehive State is experimenting with the automation of prescription renewals under the sandbox regulatory framework. The idea is that when routine tasks like this are automated, physicians will have more time to focus on higher-impact care. In accordance with Jevons’ paradox, labor becomes more valuable with automation, not less.
The story of humanity is the story of innovation. Lawmakers in Utah should continue championing innovation instead of letting speculative fears impede progress.
