One of the ways Pew Research has been gauging American attitudes about AI for the past few years is by asking this question that measures peopleâs excitement against their concern. They started asking this in 2021, in the pre-ChatGPT days, and in the latest survey six years later, Americans in general are less excited than everâand also tied for the most concerned ever.

But one group is trending the opposite direction from the rest: people in the 50-64 age range, a group that mostlyâthough not entirelyâoverlaps with what we call Generation-X, which by my math is people who are 46-61.

According to Pewâs latest findings, the majority of Americans are more concerned than excited, at 52%. Pewâs 2021 survey that used the same question found that 37% were more concerned than excited. The standout finding in the latest study according to Pew is that among adults under 30 years-old, 55% are more concerned than excitedâthe first time the youngs have been so pessimistic in this survey, but other, related surveys of young people have produced similar findings lately.

But itâs the 50- to 64-year-olds I find the most interesting. They were the only group a majority of whom didnât give the âmore concerned than excitedâ responseâthough to be clear, only 8% gave the âmore excited than concernedâ answer, so letâs not act like this group wants to kiss a chatbot or something.

Whatâs more, fewer and fewer respondents in this age group have been answering âmore concerned than excitedâ in each of the past four surveys. It was 59% in 2023 (the second highest among the groups), 56% in 2024, and 52% in 2025. Â

Oddly, among 50- to 64-year-old respondents, Pew notes that 69% believed there would be âfewer jobsâ over the next 20 years in 2024, and that percentage is actually up in 2026. Itâs now 72%.

Itâs worth keeping in mind that according to Pewâs methodology this isnât a single group of people being asked the same questions over and over. There are new recruits each year. Even if it were the same people, Pew is labeling them by age at the time theyâre being surveyed, so the same person could theoretically have been in last yearâs group of 30- to 49-year-olds, and this yearâs group of 50- to 64-year-olds.

And even though the trend in this groupâs answers over time is a little out of step with the rest of the age groups, their answers arenât massively different from the rest when taken as a whole. Still, itâs a puzzling quirk worth exploring.