The Fed is now factoring AI into interest rate decisions, marking a major policy shift

Top officials at the Federal Reserve are now taking into account how artificial intelligence might boost worker output as they map out future economic conditions, a shift that could reshape how the central bank approaches interest rates and employment targets.

Fed Chair Jerome Powell touched on this issue during his press briefing in December. Looking back at previous technology waves, he noted that past innovations brought more jobs and better pay despite initial concerns. However, he said the outcome with AI remains unknown.

Research suggests these new AI tools, especially those that generate content, could significantly change both how much workers produce and the makeup of the job market itself. Two economists studying this issue created different models to see what might happen. Ping Wang from Washington University in St. Louis and Tsz-Nga Wong from the Richmond Federal Reserve Bank mapped out several possibilities.

Dramatic productivity scenarios take shape

Their most dramatic scenario assumes AI reaches full development over several decades. Under those conditions, about 23% of workers would lose their jobs, but those still working would produce three to four times more than they do now.

Wang explained that over the coming ten years, output per worker might climb roughly 7 percent each year. He stressed this represents one possible path, not a guaranteed outcome. The technology’s ability to learn and improve as people use it drives these potential gains, he said. Workers can also figure out better ways to use AI and customize it for their specific needs, leading to major productivity jumps.

These changes could affect how the Fed pursues its two main goals: keeping people employed and preventing prices from rising too fast. In December, the committee that sets interest rates predicted the benchmark rate would settle around 3 percent in the long term. Economists at the Cleveland Fed said this would be somewhat loose compared to a neutral rate of 3.7 percent.

Data center investment boom draws comparisons to 1990s

Some market watchers see parallels between today’s rush to build data centers and the 1990s spending spree on network equipment. Dan Tolomay, who oversees investments at Trust Company of the South, said rising valuations make him more cautious about returns going forward.

Vice Chair Philip Jefferson addressed college students in Germany this November about AI’s rapid spread. He said ChatGPT now has 800 million people using it each week, up from 500 million at the end of March. A recent study found that 45.9 percent of American workers were using generative AI at their jobs by June and July, compared to 30.1 percent last December. About one-third of those who adopted the tools use them daily.

The study showed AI use is highest among younger workers with more education and higher pay. Those workers saw substantial productivity increases when they used the tools.

Research by Stanford economist Erik Brynjolfsson found AI tools helped customer support workers resolve 14 percent more issues per hour. The gains were even larger for newer employees with less experience.

In scientific work, an AI system called AlphaFold changed protein research dramatically. Five years ago, scientists understood the structures of only 17 percent of the roughly 20,000 proteins in human bodies. Each discovery took months or years and could cost tens of thousands of dollars. AlphaFold now predicts structures for all human proteins plus 200 million more.

AI could match historic innovations

Jefferson said AI might prove as transformative as the printing press, steam engine, or internet. But he cautioned that figuring out AI’s effect on jobs and prices remains difficult. While the technology could replace some workers, it might also create new job types and boost overall economic growth.

On prices, Jefferson said increased productivity could lower costs and reduce inflation. But AI could also push certain prices higher as companies compete for skilled workers and build energy-hungry data centers.

If you're reading this, you’re already ahead. Stay there with our newsletter.