AIAcademy · AIAcademy · 2026-05-16
Two surveys, read together, answer "who uses AI at work" better than either does alone.
Pew, October 2025. About 1 in 5 US workers now use AI in their job — 20% have used it at least sometimes, up from 16% a year earlier. The headline number is small but the rate of change is not: a quarter-on-quarter slope that, if it holds, puts adoption past 30% by mid-2027. The same Pew series shows the public's view of AI hardening — concern is rising faster than enthusiasm, especially among workers in AI-exposed occupations.
Brookings, December 2025. The nationwide adoption survey adds the demographic detail Pew leaves on the table. Use is concentrated by age (under-30s lead, over-55s trail by roughly 3:1), by education (college-degree workers use AI at twice the rate of non-degree workers), by income tier (top quartile leads the bottom by a similar margin), and by industry (information, professional services, and finance well ahead; healthcare, construction, and manufacturing well behind).
The shape of the gap. The two papers paired tell you AI literacy is not reaching workplaces uniformly. It's reaching the workers who were already best positioned to capture upside — younger, educated, higher-income, in industries already digitised. The Brookings team flags the implication: without intentional intervention, AI use will widen existing skill and wage gaps, not close them.
What's missing. Neither survey distinguishes casual use ("I asked ChatGPT to rewrite an email") from integrated use ("AI tools are in my daily workflow"). The 20% number conflates both. The honest read is that some form of AI now touches one in five US jobs; how deeply it touches them is the next question both teams are working on.