Back HexScope Lens

"Our Customers Skew Older"? AI Chatbot Use Nearly Doubled Among Americans 65+

POINT Key points
  • US adults using AI chatbots: 33% in 2024, 49% in 2026
  • 18-month growth: 1.7x for ages 50-64, 1.2x for 18-29

“Our Customers Skew A Bit Older”

Every time an “AI” project comes up at work, somebody says it.

“Our customers skew a bit older.”

I’ve said it too. Last fall, that one line was enough for me to knock a project down a notch on my list. Then I saw a set of US numbers published this June, and the premise behind that line started to look shaky.

Over about a year and a half, the share using AI chatbots grew 1.7x among people aged 50 to 64 and nearly doubled among people 65 and older. The numbers come from Pew Research Center, a nonprofit public-opinion research organization in the US, which surveyed 5,119 US adults using a probability sample (ie respondents drawn at random, so the group can stand in for US adults as a whole rather than for whoever chose to answer).

That premise might be out of date.

How Many People Use AI Chatbots?

In the US, about half of adults now do: the share of adults using an AI chatbot went from 33% in August 2024 to 49% in February 2026, up 16 points in about 18 months. That works out to about one in two American adults using an AI chatbot in some form.

Asked about ChatGPT alone, the share went from 18% in 2023 to 44% in 2026. Until 2025, though, Pew asked only about ChatGPT, so those years don’t join up with the all-chatbot figures.

You might read that as the young pulling the average up. But 49% is everyone, all ages lumped together, and splitting it by age shows a different picture.

Past 50, The Growth Is Faster

Between August 2024 and February 2026, the share using AI chatbots moved like this in each age group:

  • 18 to 29: 55% → 66%
  • 30 to 49: 41% → 61%
  • 50 to 64: 25% → 42%
  • 65 and older: 12% → 23%

On the raw numbers, the young are still ahead. Look at how fast each group grew, though.

In points, the gap between 18 to 29 and 65-plus didn’t narrow at all: 43 points in 2024, 43 points in 2026. On growth rate, though, it ran the other way: ages 50 to 64 grew 1.7x, and ages 65 and older about 1.9x, both faster than the 1.2x for ages 18 to 29.

So the picture isn’t “the young are surging and older adults are being left behind.” Every age group moved up, and the older groups weren’t the slow ones.

One In Four Uses One Every Day

Having used a chatbot and using one as a daily habit are two different measures, and they’re worth keeping apart.

Across all US adults, 24% use an AI chatbot every day. That includes 12% who use one several times a day and 4% who say they use one “almost constantly.” By age, daily use was 31% for 18 to 29, 34% for 30 to 49, 19% for 50 to 64 and 7% for 65 and older.

Daily use still shows an age gap. Even so, people for whom chatbots are a habit rather than something they tried once clearly exist outside the youngest group too.

The same survey also asked about confidence by age, but only in 2026, so there’s no trend to read. The share who said they were extremely or very confident in their own ability to use AI chatbots was 31% for ages 18 to 29 and 6% for ages 65 and older. That question is about confidence in using the tools, not about trusting their answers.

The base is everyone, users or not, and 77% of people 65 and older don’t use chatbots at all. Recalculating among users only (my arithmetic, not Pew’s), it’s roughly half of users aged 18 to 29 and about one in four users aged 65 and older. So I’d treat starting to use and using with confidence as separate axes.

Can You Apply These Numbers To Your Own Market?

Not directly. Everything above comes from one survey of 5,119 US adults, so if you sell outside the US, these aren’t your market’s numbers.

And wherever you sell, they aren’t an age breakdown of your own customers, so mapping them straight onto your customer base would be jumping the gun.

It’s also a self-reported survey. These are people’s answers, not a tally of actual usage logs.

“Have used” and “use every day” are kept as separate measures here, too. The 2024 figures came from slightly different wording, asked only of people who had heard at least a little about chatbots. Pew still reports them as a share of all adults, counting those who had never heard of chatbots as non-users.

Pew’s methodology page puts the margin of sampling error at plus or minus 1.6 points for the full sample and 2.6 to 4.3 points for the age groups.

Still, at least in the US, the premise that AI chatbot use is concentrated among the young holds up less well than it did in 2024. Whether your own data shows the same pattern is something you can’t know until you check.

The first thing that came to mind when I saw these numbers was my own company’s customer list. I’m not about to lay the US on top of my market as if they matched. But I can’t find a reason to skip checking anymore, either.

Hold Off On “They’re Older” Until You’ve Looked At The Ages

Taking the US data at face value and deciding your market must be the same would be risky. Carrying on with “our customers are older, so AI isn’t our concern” without ever testing it has become just as risky.

The first thing within reach is to pull the age breakdown of your customers or members from your own data. After that, check how your brand looks to AI, whatever the age group. I’d get both done this month.

If your customers lean toward their fifties and sixties and up, finding out how that group actually uses AI chatbots is clearly worth more now than it used to be.

Age used to be where the discussion stopped. I think it’s now where you start looking, group by group, at what people actually do.

Sources

Share this article
Bluesky X
Back to all articles