How Do American Voters Feel About AI?
By: Avery James and Jack Elbaum
Over the last few years, artificial intelligence (AI) has taken off as a general work and consumer product, with capital expenditures not seen in decades and fierce competition among technology firms. And though Americans still harbor concerns about what AI might mean for our future, American voters are increasingly feeling hopeful about the power of AI as well.
Comparing Echelon Insights’ own polling of American voters from August 2025 to the results of identical questions asked in July 2023, we find the voting public is becoming more optimistic about AI over the last two years. This is the case when one looks at the words people use to describe their feelings about AI, whether they think it will be helpful or harmful to humanity, and whether AI will make life better or worse.
Since 2023, the share of voters saying they are worried or scared about AI fell from 49% to 40%, and 29% to 22%, respectively. The combined share of voters saying they feel hopeful or excited rose.
However, older voters remain more worried and scared about AI than younger voters, who are generally more likely to say they are excited or intrigued by the technology. This enthusiasm among younger voters might have to do with the significant difference in AI usage across age, which we found in our July 2025 Consumer Omnibus asking about AI product use. Large fractions of American adults ages 18-49 said they used AI platforms like OpenAI’s ChatGPT (56%), Google Gemini (40%) and Meta AI (33%) in the last 12 months. Adults over 50 were far less likely to say the same (25%, 21%, and 13% respectively.)
These findings are also reflected in other public polls like Gallup’s June 2025 data finding that adults 18 to 49 are far more likely to report using generative AI tools in their regular work and life than seniors. A similar age gap in feelings about AI appears in YouGov’s June 2025 data, which found a similar young adult skew in reported excitement about AI.
This same increasing optimism, but large age gap, is apparent when it comes to our questions about whether artificial intelligence will be helpful or harmful for humanity and whether it will make one’s life better or worse. Since 2023, voters overall went from saying AI will be harmful for humanity by a three-point margin to saying it will be helpful by a 10-point margin. They shifted from saying it will make their life better by a two-point margin to a 17-point margin.
Meanwhile, voters under 50 soared from 42-33 to 55-29 in favor of saying it will be helpful to humanity over the last two years. They also moved dramatically from 35-23 to 52-18 in favor of saying it will make their lives better.
Older voters tempered their worries over the last two years as well, now reaching a statistical tie between the two descriptors. But this shift is not nearly as pronounced as that among young voters. Of course, a significant share of voters is still unsure, especially among older voters. This leaves younger voters more optimistic, which is also corroborated by YouGov’s June 2025 data finding that younger adults predict a more positive effect of AI on society.
The gap in optimism about AI also notably contains an education component. Voters with a college degree became significantly more likely to say AI will be helpful for humanity, with 52% saying it will be helpful, compared to 28% saying it will be harmful. This optimism sharply rose since our July 2023 survey, but only in this educational cohort. Voters without college degrees did not change their mixed opinions about AI since then, and a larger share remain unsure.
Despite the general shift in optimism about AI — driven by younger and more college-educated Americans — voters are not less averse to regulation. Rather, support for regulation of AI has risen since 2023, to a 59-27 majority. Most notably, the data suggest this shift in regulation attitudes is steady across age differences, as both older and younger voters increased their support for regulation ahead of concerns it could do more harm than good for a new industry. This may be because AI is perceived as less of a new industry to voters than it was two years ago.
The growth of support for AI regulation comes most significantly from Republican voters, who now say, by a 57-31 margin, that government regulation of AI is necessary to protect the public interest. A smaller shift among Independents in favor of regulation is also visible.
The data suggest American voters are becoming more optimistic about AI’s ability to help people and have more positive feelings about the technology. But this does not translate to a hands-off approach to regulation. As AI technology continues to play a larger role in people’s work and life, the voting public becomes more open to government ensuring it serves the public interest instead of protecting a new industry from the potential harms of regulation.
<<<Download the August crosstabs here>>>
<<<Download the August toplines here>>>
<<<Download the July consumer crosstabs here>>>
<<<Download the July consumer toplines here>>>