The AI Generation Gap Isn't Where You Think It Is

Australian data point the same way on use (more than 70% of under-35s have used generative AI, against 17.2% of those aged 65–74) but not on mood: younger Australians are the most positive. See the Australia section.

Two notes on method. All figures are for US adults unless stated. Pew reports age bands, not generations, so this article uses age bands and treats generation names as a loose lens; the 50–64 band, for example, mixes late Gen X and early Baby Boomers.

ChatGPT use among US adults rose from 18% in 2023 to 44% in 2026, and every age band grew (Pew appendix).

The fastest growth came from the middle, not the youngest. Adults aged 30–49 added 34 points, ahead of 18–29 (28), 50–64 (24) and 65+ (15). Pew's June 2026 survey reports the same pattern for any AI chatbot (ChatGPT, Gemini, Copilot and others):

The catch-up is real in ratio terms but not in points. In 2023, adults 18–29 were about eight times as likely as those 65+ to use ChatGPT (33% vs 4%), a gap of 29 points. In 2026 they are about three times as likely (61% vs 19%), but the gap has widened to 42 points. Both figures are calculated from the Pew appendix.

Daily and workplace chatbot use are nearly identical for ages 18–29 and 30–49, then fall sharply from 50 (Pew, June 2026).

The two under-50 bands are within 3 points of each other on daily use and 2 points on work use. The step from ages 30–49 to 50–64 is 15 points on daily use and 14 on work use (both calculated from Pew's figures). Ages 30–49 are nominally ahead, but gaps this small are better read as a tie than as a lead for the middle. The finding that matters is that the youngest adults are not ahead of their older colleagues.

Pew defines daily use as "several times a day or more" plus "about once a day". Across all US adults the daily figure is 24%.

About two in three US teens (64%) use AI chatbots, close to the 66% of ages 18–29 in Pew's February 2026 adult survey. The two surveys differ in date and wording, so treat the comparison as a rough one.

Schoolwork is the most common of the uses Pew reports, and emotional support is a minority behaviour (12%, with majorities saying they have not used chatbots for emotional support or for casual conversation). For comparison, 10% of all US adults say they have used a chatbot for emotional support (Pew, June 2026).

Both teen reports come from one Ipsos online survey of 1,458 US teens, fielded 25 September to 9 October 2025 with a margin of error of ±3.3 points.

ChatGPT is the most used chatbot in every age band, and Gemini, Copilot, Meta AI and Grok all peak at ages 30–49 rather than 18–29. Adults under 50 are about twice as likely as those 50 and over to use ChatGPT (57% vs 28%, Pew, June 2026).

The clearest age skew is at the specialist end. Character.ai is used by 7% of 18–29s and 1% of those 50 and over.

Between 14% and 18% of every age band expect AI's impact on society over the next 20 years to be positive (Pew, June 2026).

The youngest adults are the most negative (48% vs 35–39% in the older bands), even though they use chatbots heavily. Pew's own summary is that younger Americans "are not more favorable" about AI despite their higher use (Pew, Americans and AI 2026). Uncertainty runs the other way, rising from 8% at ages 18–29 to 21% at 65+.

People are more positive about AI's effect on themselves than on society, in every age band:

Published primary surveys say much more about how often and how widely people use AI than about what for, broken down by age. What they do show:

• Teens: 54% use chatbots for help with schoolwork, 16% for casual conversation and 12% for emotional support or advice (Pew, Feb 2026).
• Adults, overall: 10% have used a chatbot for emotional support (Pew, June 2026).
• Work: 43% of employed adults aged 30–49 and 41% of those aged 18–29 use chatbots for tasks at work, against 29% at 50–64 and 18% at 65+ (Pew, June 2026). A separate global survey found 74% of Gen Z and 74% of millennials use AI "to some extent" in their day-to-day work (Deloitte, 2026). The two surveys measure different things and should not be compared number for number.

Many popular articles describe a tidy use case for each generation, such as "Boomers use AI to learn about hobbies". Those claims trace back to secondary summaries rather than published survey tables, so this article leaves them out.

Gallup finds that the share of US employees using AI at work rose from 21% in May 2023 to 52% in May 2026, and daily use from 4% to 15% (Gallup).

Gallup's page does not break these results down by age, so this trend describes the whole workforce rather than any one generation.

The generational point on unsanctioned use comes from Microsoft and LinkedIn's 2024 Work Trend Index, which found bring-your-own-AI above 70% in every age group globally. The Australian results, from a survey of knowledge workers, put generative AI use above the global average:

Two Australian surveys show the same steep age gradient in use as the US data: roughly seven in ten adults under 35 have used generative AI, against fewer than one in five of those aged 65–74.

The two surveys measure different things ("ever used" and "recently used"), so they are shown side by side and not combined. Neither reports use rates for ages 35–64 in the material reviewed, which means the "cliff at 50" cannot be tested for Australia.

Attitudes tell a different story from the US figures above. In the ANU survey, younger Australians are markedly more positive about generative AI and less worried about its political use:

At least a moderate understanding of generative AI falls from 48.2% of 18–24s to 10.1% of those aged 75 and over (ANU). In the US, optimism about AI's impact on society is flat at 14–18% across age bands, while in Australia a positive view runs from 49.0% under 35 to 21.2% at 55+. The questions differ (a positive view of generative AI now, against the expected impact of AI on society over 20 years), so the numbers should not be set against each other directly, but the pattern suggests the US finding of age-flat optimism may not carry over.

Among employed Australians who use generative AI, 9.8% use it every day at work and 23.4% most days (all ages, ANU). A separate global study led by the University of Melbourne with KPMG (48,340 people in 47 countries, November 2024 to January 2025) found 50% of Australians use AI regularly and 36% are willing to trust it (KPMG Australia). The release gives no age breakdown for Australia and no Australian sample size.

The survey figures above are sourced. The implications below are Groovia's interpretation of them, not survey findings.

The figures come from nine primary sources, listed below with their sample sizes, and carry several limits a reader should keep in mind.

• Mostly US data. The Australian section draws on ANU and Google, the Australian Digital Inclusion Index and the Melbourne/KPMG study; the Deloitte and Microsoft/LinkedIn figures are international.
• Age bands, not generations. Pew reports 18–29, 30–49, 50–64 and 65+. A snapshot like this cannot separate an age effect (people use less as they age) from a cohort effect (each generation is different).
• "Ever use" is a low bar. It counts anyone who has tried a tool, so it overstates regular use. Daily use is the better measure of habit.
• Wording and dates differ across sources. Pew's appendix flags a change in question wording in 2026, so the 2025–2026 step in the trend chart should be read with care.
• Gallup has no age breakdown. Its workforce series shows the overall trend only. Its table also lists an August 2026 reading (53% total), which is left out here because the page's own dating is inconsistent.
• Not every gap is meaningful. Pew notes that not all differences are statistically significant. Gaps of a few points, such as 34% vs 31% daily use, are read here as "similar", not as a lead.