Glonce

Epoch AI survey finds daily AI use among US adults doubled in six months

researchThe Glonce editorsThe Decoder🔥66Glonce rating66Impact70Novelty40Relevance100Surprise50Players25Credibility70
Epoch AI survey finds daily AI use among US adults doubled in six months

Two representative surveys conducted by research institute Epoch AI together with polling firm Ipsos show that daily AI use among US adults has more than doubled in six months. The share of US adults who used AI on at least six of seven days rose from 8 percent in March 2026 to 19 percent in August 2026. Over the same period, the share of occasional users, those who used AI only one day a week, fell from 17 percent to 10 percent.

The March 2026 survey drew on 2,017 participants, while the August 2026 survey drew on 1,016; both samples were randomly selected and weighted to reflect the US population.

Epoch AI cautions that the two waves are not fully comparable. In March, participants were asked how often they used AI overall. In August, they were asked about each AI service separately, and researchers recorded the highest reported frequency across those services. The researchers say this difference in methodology could cause the August figures to underestimate actual usage.

Key facts

  • The share of US adults using AI on at least six of seven days rose from 8 percent in March 2026 to 19 percent in August 2026.
  • The share of occasional users, who used AI only one day a week, fell from 17 percent to 10 percent over the same period.
  • The March survey sampled 2,017 people and the August survey sampled 1,016; both were randomly selected and weighted to reflect the US population.
  • Epoch AI says the two waves are not fully comparable: March asked about AI use overall, while August asked about each service separately and used the highest reported frequency.
  • The researchers say this methodology shift could cause the August figures to underestimate actual usage.

Why it matters

The result documents a sharp shift in how habitual AI use has become among the general US population, based on survey data rather than a single company's own usage metrics. A rise from 8 percent to 19 percent of adults using AI almost daily, in just six months, suggests AI tools are moving from occasional experimentation toward a regular daily habit for a growing share of the country, while the share of once-a-week occasional users shrank from 17 percent to 10 percent over the same stretch.

Who it affects

The figures describe the general US adult population surveyed by Epoch AI and Ipsos, not the user base of any single AI company. They matter to AI companies and investors tracking how fast consumer habits are shifting, to researchers and journalists following AI adoption, and to policymakers weighing how embedded AI tools have already become in daily life.

How to use it

Anyone citing these numbers should note that the March and August surveys did not ask the same question: March asked about AI use in general, while August asked about each AI service separately and took the highest reported frequency across services. The two waves are best read as directional evidence of a large increase, not as a strictly matched comparison of two identical measurements.

How solid is it

The surveys rest on randomized, population-weighted samples of meaningful size, 2,017 respondents in March and 1,016 in August, run by a research institute (Epoch AI) with a professional polling firm (Ipsos). The source does not state a margin of error or confidence interval for either wave, and the question format changed between waves, a limitation Epoch AI itself flags rather than one this account is adding.

Risks and caveats

Epoch AI says that switching to per-service questions in August, then taking the highest reported frequency, could cause the August figures to underestimate actual usage relative to March's single overall question. The source gives no breakdown by age, region or other demographic group, does not name which AI services respondents were asked about in August, and does not explain why the question format changed between the two waves.