This polling was conducted by Epoch AI in partnership with Ipsos, using the KnowledgePanel, a probability-based online panel recruited via address-based sampling. Probability-based sampling reduces self-selection bias and produces estimates that are more representative of US adults.

The survey was fielded October 9–11, 2026, and includes 2,015 respondents. Some questions were asked only of employed respondents (respondents working full time or part time, from the Ipsos panel profile), some only of respondents who had used an AI service in the past 7 days, and some only of respondents who met both conditions.

At the start of the survey, respondents were told that “AI service” means consumer-facing generative AI products such as ChatGPT, Claude, Google Gemini, and Microsoft Copilot. We defined “personal use” as any activity that is not part of a paid job. We defined “work” as any activity done to earn money, including a full-time or part-time job, freelance or contract work, or running one’s own business, whether formal or informal.

This is the first wave of a monthly tracker that will ask the same core questions each wave. Several questions differ from our polls conducted in March and July 2026: questions now ask about the “past 7 days” instead of the “past week”, the scale for days of use now stops at “5 or more days”, and the question about who paid for the AI services used for work is now one overall question instead of one question per service. Changes in question wording may affect how respondents report their usage across waves.

The full text of each option is available by clicking on the bars in the graphs above.

The results have been weighted to represent the adult population of the United States, ages 18 and older, using US Census benchmarks. The service lists included a fictitious service to identify inattentive respondents. Respondents who select it are removed before weighting. No respondent selected it.

Respondents who were asked a question but did not answer it are included in its base and reported as “No answer” in the downloadable data. They are not shown in the graphs.

The graphs show 90% Wilson score confidence intervals, computed on an effective sample size that accounts for unequal weighting. For each weighted proportion \(\hat{p}\), we compute

\(\frac{1}{1 + z^2 / n_\text{eff}} \left( \hat{p} + \frac{z^2}{2 n_\text{eff}} \pm z \sqrt{ \frac{\hat{p} (1 - \hat{p})}{n_\text{eff}} + \frac{z^2}{4 n_\text{eff}^2} } \right)\)

where \(z = 1.645\) is the 90% two-sided critical value from the standard normal distribution, and \(n_\text{eff}\) is the Kish effective sample size,

\(n_\text{eff} = \frac{\left( \sum_i w_i \right)^2}{\sum_i w_i^2}\)

computed over the respondents in the relevant base. Unlike the normal-approximation interval, the Wilson interval always stays within 0–100% and remains informative for proportions near 0% or 100%.

All estimates are weighted. The estimated proportion for any response category is:

\(\hat{p} = \frac{\sum_i w_i y_i}{\sum_i w_i}\)

where \(w_i\) is the weight for respondent \(i\) and \(y_i\) is a binary indicator (1 if the respondent selected the category, 0 otherwise).

Standard errors, reported in the downloadable data, are computed using Taylor series linearization to account for weighting. The Ipsos KnowledgePanel uses an unclustered and unstratified probability-based sampling design. The variance estimator therefore simplifies to:

\(\hat{V}(\hat{p}) = \frac{n}{n - 1} \cdot \frac{1}{(\sum_i w_i)^2} \cdot \sum_i \left[ w_i (y_i - \hat{p}) \right]^2\)

where \(n\) is the number of respondents in the relevant base.