CORRECTS-MNI INTERVIEW: Fed Survey Sees Powerful AI Boost

(Corrects second paragraph to make clear that 60% of firms invested in AI in 2025, not 2026)

A new Fed survey of CFOs suggests the AI revolution is outpacing the 1990s tech boom, offering Federal Reserve officials early evidence that a productivity surge may be closer than data show and the overall workforce impact more modest than feared, Atlanta Fed economist Brent Meyer told MNI.

The survey of roughly 750 senior financial officers in the Atlanta and Richmond Fed's CFO panel and through Nasdaq and Duke University networks finds that 60% of firms invested in AI in 2025 and over 80% expect to invest this year. Larger firms are investing more than smaller firms, and there is little evidence that AI usage has affected or will affect the total number of employees. 

"We're on the cusp of a dramatic increase in AI adoption, and it looks like it's happening at least as quick, if not faster, than the technology revolution in the '90s," said Meyer, co-author of research on the survey. "If these expectations for the impact on sales revenue and employment are correct, then the implied productivity gains are pretty meaningful."

Firms with high-skill services like finance saw "sizable" annual labor productivity growth of about 0.8% last year, while low-skilled services and manufacturers saw roughly half of that, Meyer and the co-authors calculate. These effects are expected
to roughly double in 2026, with the largest anticipated increases in high-skill services and finance at over 2%.

"There's a sizable uplift in productivity and there's a differential in the way that gains are coming through to different size firms and sectors," Meyer said. 

PICKUP IN PRODUCTIVITY

The findings come as Fed officials have already begun nudging up their long-run GDP estimates on a post-Covid climb in productivity. In the most recent Summary of Economic Projections, the median longer-run growth estimate rose by two-tenths, with Chair Jerome Powell citing recent efficiency gains as firms worked around labor shortages during the pandemic. (See MNI POLICY: AI Boom Complicates Fed's Path To Lower Rates)

The AI productivity boom should build on that, Meyer said, adding that he sees echoes of the Solow paradox as CFOs' perceptions of productivity gains appear to be running ahead of what the hard data capture. 

"If what we're measuring within the firms in our survey applies to the entire economy, we're likely to see a pickup in productivity growth going forward -- a pretty meaningful one," he said. "It may be five, ten years down the line before we really see it. But firms are telling us it's going to happen." 

And unlike the technology buildout of the 1990s, which required substantial capital investment in hardware and network infrastructure, the productivity gains from AI appear to be flowing primarily through total factor productivity -- the so-called Solow residual -- rather than capital deepening, Meyer said.

"For the typical firm, they're reaping the benefit of AI without having to invest a lot in equipment," he said. That may blunt concerns that smaller firms will fall behind their larger rivals. "Because it's relatively easy compared to the past to engage with this new technology, there may not be as big a first-mover advantage this time around." 

Larger firms are nonetheless spending heavily on bespoke AI systems. The median large adopting firm in the survey spent approximately USD300,000 on AI in 2025.

JOBS FOR CREATIVE WORKERS

The survey suggests AI is not the driver of the recent cooling in labor market churn, at least not yet, Meyer said. 

CFOs identified their primary motivation for AI investment as boosting production efficiency and enhancing decision-making rather than cutting headcount -- a finding that would assuage fears of mass layoffs. "It's much less about reducing costs or downsizing employment," Meyer said.

Projected out to 2028, the researchers found the dominant employment trend to be a shift in the mix of workers rather than a net reduction. Administrative staff, accountants, financial analysts and software developers are among the roles the survey identifies as most exposed to AI displacement. In their place, firms anticipate growing demand for higher-skill, "cognitively intensive" workers. 

"The overall message from what we found on the employment side is very, very modest effects in 2025, and modest but growing effects in 2026. "Further out in 2028, things that are routine -- those jobs are likely to get replaced with more high-skill, productive, creative class individuals. We see a reconfiguration of the mix of employees in a firm."