Using artificial intelligence might trigger childhood memories of your parent telling you to “just try it” when a new dish landed on the supper table.

Well, the Phoenix Center for Advanced Legal & Economic Public Policy Studies found that professional adults might be having flashbacks; this time about AI in their workplaces.

According to its latest research project, the non-profit organization that studies broad public-policy issues related to governance, social and economic conditions, with a particular emphasis on the law and economics of the digital age, said most American workers believe artificial intelligence will reshape the economy.

However, Phoenix Center chief economist George Ford also found far fewer professionals believe AI will reshape their own jobs.

In a new analysis released on Monday entitled “Substance or Spectacle? AI Use and the Personal–Societal Risk Wedge,” Ford explained that the workers best positioned to dismiss AI as over-hyped are instead the ones most likely to conclude the technology is coming for their own work.

Drawing on Wave 119 of the Pew Research Center’s American Trends Panel, Ford documented what he called “a striking wedge” in how workers assess AI.

Ford’s effort determined nearly three in four expect a major impact on workers generally, but only about one in three expect a major impact on themselves

“On its own, that gap is ambiguous: it could reflect workers who genuinely face lower risk, or workers comforting themselves,” the Phoenix Center said in a news release.

Ford pointed out that what the data can identify is how the translation of societal concern into personal concern shifts as workers gain firsthand experience with the technology.

Ford tested three competing explanations, each making an opposite prediction. They included:

—If proximity makes personal stakes concrete (salience), then heavy users should translate societal concern into personal concern more

—If experience reassures (motivated reasoning), then they should translate it less

—If the self gap is a fixed disposition, then experience should make no difference

Ford said the data delivered a “clear verdict.”

The Phoenix Center highlighted heavy AI users convert general concern into personal concern at a markedly higher rate than light users’ a gap of roughly 10 percentage points that holds up whether or not the analysis accounts for the demographic and economic traits distinguishing the two groups, and that no standard trait reproduces.

“The result bears the signature of salience: familiarity does not breed comfort, it brings the matter home,” the organization said.

Ford went on to note that the finding carries a pointed policy implication.

Ford explained if the gap between societal and personal concern closes with use rather than disposition, then today’s relative calm among most workers is unlikely to last as AI diffuses through the workplace.

He added that’s a consideration for policymakers weighing the pace and scale of workforce and displacement responses.

“The workers who have actually worked with AI are not talking themselves out of the risk. they are talking themselves into it,” Ford said in a news release. “The comfortable distance most workers keep between ‘AI will change everyone’s work’ and ‘AI will change mine’ looks like a lagging indicator, held in place by inexperience rather than by any durable fact about their exposure.

“The workers who already see it arriving at their own desks are not anomalies. They are the rest of the workforce, just a bit early,” he added.

The complete study can be found here.