Boarding Group A… or Not Flying at All? Who Is and Isn’t Flying
March 13, 2026
Everyone’s Using Generative AI…Right?
March 23, 2026
Since the launch of ChatGPT, which launched the era of generative AI and large language models, we have been tracking consumers’ familiarity and usage of the various platforms. We continued to track this in our most recent wave of research-on-research. In part one of our two-part series, we dive into updated data on the familiarity of generative AI platforms.
Let’s jump in.
Overall
In our latest wave, ChatGPT continues to reign supreme in familiarity, with 38% of respondents saying they are very or extremely familiar with it. Gemini follows at 33%, while Copilot reaches 26%. Familiarity with the remaining platforms is lower, with Claude and Grok each at 17%, Perplexity at 16%, and Duet at 13%. The results show a clear ranking among platforms. ChatGPT leads by several points, with Gemini and Copilot forming a second tier of familiarity. The remaining platforms cluster in the mid‑to‑high teens, indicating that recognition exists but is less widespread than for the leading tools.
Wave-Over-Wave
Look at familiarity levels across the last four waves of our research-on-research; it generally increases through the earlier waves before leveling off or dipping slightly in the most recent wave. ChatGPT rises from 30% in Wave 1 to 35% in Wave 2 and 40% in Wave 3 before easing slightly to 38% in Wave 4. Gemini follows a similar trajectory, climbing from 24% to 26%, then to 33%, where it remains in the latest wave.
Copilot increases from 20% in Wave 1 to 26% by Wave 3 and holds at that level in Wave 4. Grok also grows steadily from 11% to 17% by Wave 3 and remains there in the latest wave. Claude moves from 11% in Wave 1 to 17% in Wave 4, while Duet and Perplexity show smaller increases over time, ending at 13% and 16%, respectively. Overall, familiarity builds across the early waves before stabilizing or easing slightly in the most recent measurement.
Gender
Similar to past waves, males are more familiar than females across all generative AI platforms. The gap between the genders ranges from 4 to 13 percentage points, with the largest differences between Grok and Claude.
Age
When looking at the data by age, respondents aged 25 to 44 have the highest familiarity with each panel, except ChatGPT. Across all platforms, there is a significant drop in familiarity with respondents 45 or older.
Income
Familiarity with generative AI platforms generally increases with household income. The top three platforms, in terms of familiarity, remain the same. Respondents with a household income of $100K+ are more familiar with all platforms than those at other income levels, by between 15 and 18 percentage points.
Political Affiliation
Across political affiliations, Republicans and Democrats have similar levels of familiarity with each platform, with the largest difference being 5 percentage points (Grok). Independents have lower familiarity than the two major political parties across all platforms, by 4 to 8 percentage points.
Ethnicity
Breaking down the data by ethnicity highlights some variation in familiarity. While ChatGPT has the highest familiarity across all ethnicities, Caucasian’s familiarity is between 10 and 14 points lower than that of other ethnicities. The trend of Caucasian familiarity being lower than that of other ethnicities holds for Gemini and CoPilot, but not for Grok or Claude.
Region
Surprisingly, there were significant variations in familiarity across regions. The West has the highest levels of familiarity across all platforms, followed by the Northeast. The Midwest has the lowest level, with its familiarity between 3 and 9 points lower than other regions.
Panel
There are significant differences in levels of familiarity with generative AI. Panel E has the highest level of familiarity with all platforms, sometimes more than double that of other panels. Panel O has some of the lowest levels of familiarity, with its respondents least familiar with ChatGPT, CoPilot, and Gemini.
This data highlights how panel selection can bias results. That is why it is best practice to blend sample providers to mitigate any potential bias strategically. Click the button below to learn how EMI can help you on your next project.
