This is the biggest myth about the AI-powered economy right now, according to one veteran investment chief

• Jim Paulsen says the link between AI and productivity may be tenuous.
• Tech giants have laid off workers recently, citing productivity gains from AI.
• Paulsen argues AI's productivity boost might not bring an economic windfall.

A common narrative of the AI boom says that the technology will lead to vast gains in productivity.

Throughout 2025 and 2026, companies like Amazon, Meta, and Block have laid off chunks of their workforce and cited AI as at least part of the reason, claiming that the technology has made their workforce more efficient.

Wall Street veteran Jim Paulsen is wary of the narrative. Not only that, but he thinks that the productivity gains, if they're happening at all, might not lead to stronger economic growth.

Paulsen, the former chief investment strategist at The Leuthold Group, took a detailed look at actual job growth and productivity in the US since the dot-com boom of the early 2000s. He argued in a Substack post that this era shows that the relationship between productivity, employment gains, and the overall economy isn't clear-cut.

"During the last 25 years, the link between innovations and economic performance has so far proved disappointing," Paulsen said. "Unlike most of U.S. history, today's productivity gains — if there are any actually occurring — only happen at the expense of job losses. Moreover, the dissemination of innovations to other parts of the economy beyond the information sector have so far proved elusive."

The veteran investing pro highlighted the importance of evaluating major market trends, such as the impact of AI on workplace productivity, using the information sector as an example.

While its companies boasted a 12% productivity gain, opposite the 2% growth in the rest of the economy, the boost came only after most companies had effectively stopped hiring. Fewer staff members doing similar work looks like productivity on paper, but it doesn't show actual improved efficiency or sustainable growth.

"Since the dotcom peak, job creation outside of the information sector has been noticeably slower and information sector employment has significantly contracted," he wrote. "Indeed, it's no coincidence the single year that information sector productivity contracted was in 2022 (see chart 2) when job creation in information sector soared."

In the current economic climate, it has been easy to compare the AI boom to the early dot-com era. Paulsen noted this, acknowledging that many prominent voices in technology, markets, and economics have taken this perspective, but he doesn't think the comparison is appropriate.

"The contemporary environment is nothing like the 1990s nor even anything like U.S. history prior to the 1990s," he said. "The correlation between annual productivity growth and annual employment growth from 1960 to 1999 was +0.17 whereas since 2000 the correlation has been -0.70," he said, adding that from 1960 to 2000, heightened productivity meant higher job growth, while the opposite has been true since.

Paulsen said that he sees the current landscape as unlike any other period in recent history, making it difficult to assess how it will unfold. The last 25 years of history, however, show that the current era is different from what came before.

"Since the dotcom bust, unlike most of U.S. history, productivity results have been mostly limited to periods when overall job creation has been weak," Paulsen said.