Economists are debating how AI might reshape the U.S. job market and boost productivity growth in the decades ahead. But a new paper finds that such benefits will be harder to match across the Atlantic.

Why it matters: The result might be a widening divide in innovation and economic outcomes between the U.S. and continental Europe.

Driving the news: As it stands, the U.S. and Europe are on divergent paths for generative AI adoption, according to a new paper by the McKinsey Global Institute — a predictable result, given that most of the technology has been built in the U.S.

• "Automation technology has the potential to revive productivity, allowing economies to solve most of today's labor market challenges," the paper notes.
• "Europe and the United States are not on the same trajectory for capturing this productivity growth: most AI-related innovations are developed in the United States."
• In the past, Europe has moved more cautiously with all kinds of technological innovations, regulating them more heavily amid worries about labor-displacing technology. AI could turn out to be another example.

State of play: U.S. productivity has been booming — a development that probably doesn't have much to do with AI adoption so far. The rest of the world hasn't kept up.

• "It's just mind boggling that productivity [growth] in the United States between 2019 and now has been 6%. In Europe, 0.6%," European Central Bank president Christine Lagarde said at an event hosted by the Council on Foreign Relations last month.

The big picture: The McKinsey paper says in a scenario of widespread AI adoption, occupations that require repetitive tasks and low-skill data processing could see plummeting demand, including sales and office support.

• In that case, the paper estimates both the U.S. and Europe will need as many as 12 million people to transition to different types of jobs over the next half-decade.
• That could be an easier lift for the U.S., where transitions of this scale are in line with the pre-pandemic norm. In Europe, it's double the pre-pandemic pace.

What they're saying: "If you look at the potential for Europe, it would require a step up in that metabolic rate of people being able to move to doing different things. In the U.S. this rate has been higher historically," Anu Madgavkar, a partner at the McKinsey Institute, tells Axios.

• About 1.2% of the U.S. workforce shifts jobs each year, whereas 0.4% of the European workforce changed occupations annually between 2016 and 2019, the paper says.
• Without the required worker shifts, the paper warns of a polarized labor market: one with more higher-wage jobs than workers and too many workers for lower-paying jobs.

Go deeper

The Fed should be more verbose in its policy announcements, a top official argued today. As practitioners of the art of "Smart Brevity," we'll try not to take this personally.

What they're saying: The central bank's policy statements "have become shorter over time," Cleveland Fed president Loretta Mester said at a Bank of Japan conference, according to a prepared text. The one from the last meeting clocked in at 359 words.

• "While simpler is often seen as a virtue, it can also be a detriment, since policymaking has to be done in an uncertain world, one in which the economy is constantly being buffeted by shocks that can lead economic conditions to evolve differently than anticipated," she said.
• "With short statements, each word takes on added significance. Short statements suffer from what I call a 'Hotel California' problem: we are reluctant to change particular words because of the possible signal that doing so may send," she said.
• "Words 'check in' but it is hard to get them to 'check out' even when it is desirable," she said, such as the Fed's use of the word "transitory" to describe inflation even as price pressures were taking off in 2021.

Mester also called for the Fed to connect officials' interest rate outlook projections to their economic activity projections so that central bank watchers can better understand how economic developments are likely to affect policy decisions.