
From Prompt Curious to Prompt Power User: Who’s Actually Talking to AI?
October 7, 2025
Strategic Blending in the Age of AI: Why Automation Alone Isn’t Enough
October 9, 2025Generative AI is not just for making funny lyrics in the style of Taylor Swift; people are using it for a variety of different tasks. We wanted to understand not only if people were using the tools, but what they were being used for. In the most recent round of research-on-research, we asked respondents who indicated they had used a generative AI tool in the past 30 days what they were using the tool for.
Let’s dive into what we uncovered.
Overall
The most common uses of generative AI for respondents are writing assistance (36%), brainstorming ideas (35%), and explaining complex ideas in simple terms (35%). Other frequently cited activities include summarizing long articles (31%), generating images (32%), and planning meals or recipes (25%). Smaller shares report using AI for budgeting (17%) or medical decisions (16%), while 7% mention other unique activities.

Gender
For certain tasks, men and women use AI differently. Men are more likely to generate images (35% vs. 28%) and explain complex ideas (40% vs. 29%). They also turn more often to AI for medical-related decisions, with 19% of men versus 12% of women reporting this use.
Women, however, are less engaged with summarizing content, at 24% compared to 36% of men. Entertainment-related suggestions also reveal a difference, with 29% of men using AI for recommendations compared to 21% of women. Men tend to rely more heavily on AI for technical and decision-oriented tasks, whereas women are less likely to engage in these areas.

Age
Younger respondents have the highest usage levels across all tasks. Brainstorming ideas peaks at 43% among 18–to 24-year-olds and 41% among 25–to 34-year-olds, compared to just 19% of those 65 and older. Similarly, summarizing content is far more common among younger adults, with more than a third of those under 45 using AI in this way, while fewer than one in five older adults do. Image generation follows a similar trajectory, with usage as high as 37% among those aged 35–44, but falling to 11% among seniors. Interestingly, older respondents are more likely to select “Other,” with 20% of those 65 and above giving unique examples compared to just 3% among the youngest group. Upon reviewing the open-ended responses, many of these older users stated that they primarily use AI for general search purposes.

Income
Income strongly shapes AI usage patterns. Writing assistance rises steadily with income, from 23% among those under $20,000 to 44% among households earning $100,000 or more. Summarizing long articles follows a similar path, starting at 22% in the lowest bracket and rising to 39% at the highest. Brainstorming ideas show similar growth, from 26% to 42% across income levels. Entertainment recommendations and explanations of complex ideas also lead to higher engagement among wealthier households.

Political Affiliation
Task usage shows variation across the political spectrum. Republicans (20%) are the most likely to use AI for medical decision-making, compared to 16% of Democrats and 12% of Independents. Summarizing long articles is most common among Democrats (35%), while Independents trail at 26%. Meal and recipe planning is relatively consistent across groups, though Republicans and Democrats (28% and 27%) are slightly ahead of Independents (21%). Overall, Republicans and Democrats tend to report higher levels of usage across most activities, while Independents and Others report lower.

Ethnicity
The usage of generative AI for specific tasks also varies by ethnicity. Writing assistance is particularly common among Asian respondents (43%) and African American respondents (38%), compared to 36% of Caucasians. Summarizing long articles is highest among Hispanic/Latino respondents (39%), above other groups. Meal and recipe planning stands out among African American (30%) and Hispanic/Latino (29%) respondents, compared to 24% of Caucasians. Explaining complex ideas is more often reported by Asian respondents (44%) and those in the “Other” category (37%), compared to 36% of Caucasians.

Panel
The panel source reveals significant variation in how people utilize AI. Panel N has the highest share of users using AI for brainstorming (47%), compared to just 28% in Panel D. Similarly, 41% of Panel Q use AI for summarizing content, while Panel D has the lowest rate at 18%. These are some of the most significant gaps observed across tasks. Budgeting and financial advice shows an even wider divide, ranging from only 6% in Panel J to 27% in Panel Q. The contrast of more than 20 points between these panels represents the single most considerable difference.

These examples highlight how panel composition can shape results, making strategic blending sample essential for a more reliable view of behavior. Click the button to learn more!



