Near-daily AI use doubled. The survey redesign blurs by how much.
Samaritan Research and Ipsos estimate that 18.7% of US adults used AI on six or seven days in late August, up from 8.2% in March. The direction fits other evidence; the precise change is softened by a changed questionnaire.
AI ADOPTION · SURVEY METHOD REVIEW
The number moved fast. The share of US adults who used AI on six or seven days of the previous week went from 8.2% in early March to 18.7% in late August 2026. The confidence intervals do not overlap, and the sample is probability-based and weighted. But the survey also changed how it asked about days of use, so “more than doubled” tells you the direction, not the speed.
QUICK ANSWER
The habit is deepening faster than the audience is growing
Any weekly AI use rose 6.4 percentage points, from 50.0% to 56.4%. Near-daily use rose 10.5 points. The striking movement is inside the user base: occasional users appear to be becoming habitual users.
What the two surveys found
Samaritan Research and Ipsos fielded two surveys through Ipsos KnowledgePanel. March 3–5 included 2,017 US adults; August 28–30 included 1,016. After attention-check exclusions, estimates were weighted to represent adults 18 and older. Near-daily use rose from 8.2% (90% interval 7.2–9.3%) to 18.7% (16.7–20.8%).
| Frequency | March | August | Change |
|---|---|---|---|
| No AI use | 50.0% | 43.6% | −6.4 points |
| 1 day | 17.3% | 10.1% | −7.2 points |
| 2–5 days | 24.3% | 27.3% | +3.0 points |
| 6–7 days | 8.2% | 18.7% | +10.5 points |
Who is inside the number
Near-daily users made up roughly 16% of weekly AI users in March and 33% in August. Weekly reach grew modestly, while intensity roughly doubled among the people already in.
AI use may be shifting from discovery to routine before it reaches everyone.
Then the questionnaire changed
March asked one overall question using three broad bands. August asked an exact count for each service. Researchers took the maximum across services and mapped it back to the bands. The maximum may understate use across different services; repeated prompts may improve recall or inflate it; exact counts invite rounding. Samaritan explicitly calls cross-wave comparisons approximate.
A second opinion from Pew
Pew surveyed 5,119 US adults in February 2026 and found 24% used AI chatbots daily: 8% about once a day, 12% several times a day and 4% almost constantly. Pew measured chatbots differently, so 24% is not a replication of 18.7%. It shows that daily use in the high teens or twenties was plausible.
What a count of days cannot say
Six days of use could be a two-minute search each day or an eight-hour coding session. Frequency says nothing about the task, the duration, the value, the accuracy, the dependency or who benefits. It measures habit, not productivity.
What I would measure next
- Keep one repeated frequency question unchanged.
- Measure minutes and tasks.
- Follow the same people over time.
- Link frequency to quality, time saved, errors and confidence.
- Publish demographic and occupational cuts.
Verdict
Two probability-based surveys indicate a large rise in near-daily self-reported AI use between March and August. Confidence intervals and Pew support a real habit shift. Because the questionnaire changed, 2.28× is not a precise growth rate. The structural signal is that AI is becoming routine for a growing share of people who already use it.
FAQ
Did near-daily AI use really double?
The estimate rose from 8.2% to 18.7%. Direction is clear; the exact ratio is approximate.
How many people were surveyed?
2,017 in March and 1,016 in August.
Does frequent use prove productivity?
No. Frequency measures repetition, not value.