The loudest voices in the artificial intelligence debate are getting it wrong: AI isn’t going to radically transform the economy overnight.
The bad news about that is we aren’t going to enter a world of unlimited leisure time anytime soon. The good news is we aren’t on the precipice of a massive surge in unemployment from AI replacing workers.
What is likely to happen in the next few years is more muted and more predictable. Just as in previous cycles of technological advancement, we are entering a period of volatility. It won’t be suddenly earth-shattering.
Every new technology comes with terrifying headlines about potentially catastrophic job losses. In his seminal work on the impact of automation on jobs, MIT economist David Autor noted the tendency of headlines to overstate the extent of job losses from new technology. In 1961, Time magazine ran the doomsday headline “Business: The Automation Jobless.” The story beneath was an inflated reaction to new technology that quoted former Pennsylvania Congressman Elmer J. Holland saying “one of the greatest problems with automation is not the worker who is fired, but the worker who is not hired.” Sound familiar?
Sure, this time could always be different. But the burden of proof is on AI enthusiasts (and AI doomsdayers) to show that this time, technology really will buck historic trends. Until then, workers, business owners, and policymakers should keep a cool head.
Bleak February jobs data—92,000 jobs shed, another uptick in the unemployment rate—triggered a barrage of worrying headlines and prognostications that the labor market may finally be headed for a recession caused by AI adoption.
But economic research indicates catastrophic job loss like the kind Holland feared happens at an industry level, not as a broader, economy-wide effect. New waves of technology disrupt whole sectors. The research is also clear that while technological changes can reshape industries over decades, they have never led to a permanent, economy-wide loss in employment. The AI frenzy has overshadowed that broader truth.
Until we see persistent, continual job losses over a long period of time across the entire economy, the safe bet is that history is simply rhyming. AI will disrupt segments of the labor market. Future generations of accountants, lawyers, and economists will be disrupted most. But you shouldn’t expect your children to be lying around all day, navel-gazing, or working 12-hour shifts in a South American lithium mine.
You should, however, be concerned they will struggle to land their first white-collar job and that your company will find it difficult to hire qualified workers to oversee vast and complex AI systems in the future.
That gets at the real AI-risk we see forming in the material world in the next few years. Pressure to shift toward AI may strip out the human know-how of today’s younger workers. As they climb the ranks, they won’t have the knowledge companies will need to benefit from AI-productivity gains.
To best harness AI, humans need to know how to build, manage, and grow AI systems to align with company-specific strategies. That can only be achieved by investing in developing a well-trained workforce, including managers. Companies need to be brave in investing in their talent pipeline, which will pay dividends much further down the road. Those that meet the AI revolution by investing in their workforce will reap the greatest rewards.
The action plan for the policymakers is murkier. The next chair of the Federal Reserve can work overtime to convince his colleagues of the need to promote economic growth through lower interest rates, but there is very little that policymakers, particularly monetary policymakers, can do to offset AI-related shifts in the labor market.
Positive productivity shocks like the kind AI will likely bring should be disinflationary. That could, in theory, be a win for the Fed. Faster productivity growth could raise the economy’s capacity to withstand higher interest rates, making the Fed’s job a lot easier.
Modeling by New York Fed President John Williams and the economist Thomas Laubach has found the neutral interest rate has been rising over the past year. In fact, the latest estimate of the nominal equilibrium rate implied by their model is just over 3.75%—at the top of the current target range for the federal funds rate.
There are many potential causes for this rise that are unrelated to AI. However, one could reasonably expect an AI-related positive productivity shock to further lift the nominal equilibrium rate, affording the Fed time to wait out the latest round of supply shocks and inflationary pressure.
Unfortunately, that is a utopian scenario. The Fed can’t influence technological advancements or how those advancements spread throughout the economy. Nor can they control the size of the disinflationary boost from the rise in productivity or address the negative distributional impacts from AI-related layoffs. All they can do is keep the ship afloat and hope for the best.
About the authors: Tim Mahedy is the CEO and chief economist and Guy Berger is senior advisor, labor markets at Access/Macro.
Guest commentaries like this one are written by authors outside the Barron’s newsroom. They reflect the perspective and opinions of the authors. Submit feedback and commentary pitches to ideas@barrons.com.