Economist Robert Solow famously said of productivity gains that resulted from the PC boom of the 1980s that you could see the benefits everywhere except in the data.

This is known as the Solow Productivity Paradox.

It feels like this is the baseline assumption of how AI’s impacts on the economy will feed through. Jay Powell, for instance, said last week that “I think economic forecasters are very skeptical of … periods of high productivity, because they're so rare. And they're often revised away.”

Worker productivity, however, has been elevated for the last few years. The most recent data released on Tuesday showed worker productivity in the fourth quarter of last year rose 2.5% compared to the prior year. In the 2010s, productivity rose at more like a 1% rate.

But again, as Powell notes, current productivity bumps aren’t really capturing any AI-related contributions yet.

In a note to clients published Tuesday, economists at Barclays led by Jonathan Millar argued that “a closer look reveals that the evidence in favor of sustained acceleration is still in dispute.”

The broad strokes of Millar’s argument note that productivity can be measured in two ways — either via income or output. The two measures, essentially what a company makes off a unit of labor or what it spends to receive the output of a unit of labor, should balance.

But Millar’s team finds that these measures have “diverged meaningfully.”

In a separate note out Wednesday morning, Oliver Allen at Pantheon Macroeconomics noted that the economic impact of AI-related productivity gains isn’t a settled question, either.

Some economists — Allen clocks recent quotes from Kevin Hassett on this front — see AI as an obviously disinflationary force. The logic here being that companies getting more out of each worker see their unit labor costs fall. Sure. This is actually the story Allen mostly sees in the data right now.

But longer-term, it isn’t clear whether AI is labor-displacing or labor-augmenting, to use his terminology. How AI adoption breaks along this vector will hold the key to future inflation and productivity questions posed by this technology boom.

“For now, stable payrolls in those industries most exposed to the impact from AI suggest it is at least partly labour-augmenting, supporting wages and limiting the decline in unit labour costs,” Allen wrote. “But these are still very early days in the AI adoption process, and early evidence of its effects on the labor could easily prove temporary.”

“Moreover, even if AI dampens growth in unit labor costs, that might just boost profit margins. That depends on the currently unknowable balance between the strength of both demand and competition in those industries where AI has the biggest impact.”

Which raises another paradox investors will wrestle with in the years to come — AI might replace some workers, but put even more credentialed and expensive ones in higher demand.

Click here for in-depth analysis of the latest stock market news and events moving stock prices

Read the latest financial and business news from Yahoo Finance