Almost 4 in 5 American adults (78.5%) have used an AI tool to influence a decision they‘ve made, according to a new survey of 2,127 US adults conducted by Clear Spark Digital. The survey findings show that AI tools have moved beyond experiment and into social conventional norms to influence the decision making of most Americans.
The survey also revealed that 97.6% of Americans have used an AI tool such as ChatGPT, Gemini, or Copilot to search for information. More than 3 out of 4 respondents (76%) said they use AI tools at least weekly, and almost half (48%) use them daily. Just 2.4% have never used an AI tool to find information.
Decision-Making Spans Demographics
One of the survey’s most interesting findings is that the percentage of Americans that have made a decision based on information from AI stays largely the same across almost all demographic groups. For generations, Gen Z (respondents age 18-28), reported a 79% decision-making rate, Millennials (29-44) 79%, Gen X (45-60) 77% and Boomers and older (61+) reported 74%.
Men are slightly more likely than women to have made decisions based on AI, at 82% vs 75%. As education levels increase, so does the likelihood of using AI to influence decisions: 76% for high school graduates, 78% for technical or community college, 79% for those with an undergraduate degree, 81% for graduate degree holders, and 80% for those with a doctorate.
The uniformity in different demographic groups indicates that the use of AI in decision-making is not fueled by specific age groups, gender, or educational background. Rather, it is now commonplace for American adults.
Trust in AI Has Grown Quickly
The observed decision-making results feed into a larger story of rapid normalization of and trust in AI information, including AI overviews in search results. According to the survey, 68% of Americans believe that AI results are as trustworthy or more so than traditional Google search results. For a technology that has only been mainstream for about two years, that amount of trust was garnered quickly relative to Google‘s two-decade-long run as the default search engine.
When comparing how different generations trust AI, the findings are opposite of what many would assume about who is most likely to adopt new technology. Adults 61 and older were more likely to trust AI answers over Google search compared to Millennials and Gen Z adults. 33% of Boomers said they trust AI more than Google search results, compared to 23% of Gen Z adults who feel the same. Gen Z was also the most skeptical demographic in the survey, with nearly 4 in 10 saying they trust AI less than they trust organic search results.
In the report, Clear Spark Digital, acknowledges that the relatively small sample of older consumers (100 respondents 61 or older) means that the results of the Boomer-specific percentage of trust should be taken with some caution, but the trend of increased trust with age was consistent for every generation.
Most Americans Do Not Verify AI Information Consistently
Only 17.4% of AI users said they always verify answers provided by AI. Approximately 12% of users said that they “rarely or never” verify the output AI provides. The discrepancy in verification behavior is even more stark for those who use AI daily. Only 14% of daily users said they “always verify,” while 41% said they “verify often,” but not always.
When users do verify AI-generated answers, Google is the main resource. 72% of AI users say they do a Google search when they want to verify an AI answer. Even among those who said they trust AI more than Google, 59% still use Google to verify.
Some of the other more common ways users said they verify information from AI were: cross-reference it with trusted sources (60%), manually verify the source/citations the AI was citing (46%), and ask follow-up questions to the same AI (41%). Around 15% said they will verify information by asking a different AI tool.
The results seem to show that even though trust in AI search has grown in a relatively short period of time, the more traditional search engines still remain the place users go back to when cross referencing the information they’re given.
Verification Habits Vary by Education and Age
Education was strongly associated with verification behavior. Doctorate recipients had the highest verification levels, with 65% indicating that they always/often verify and only 4% indicating they rarely or never verify. Conversely, among those with high school degrees, 50% indicated they always/often verify and 15% indicated they rarely or never verify.
Age groups: Younger users are more likely to verify AI answers than older users. Both Gen Z and Millennials reported 57-58% always-or-often verification, while Gen X and Boomers were only 47% and 48%, respectively. The age groups that were more skeptical and hesitant to trust AI are also more likely to verify information from AI tools.
A Significant Shift in How Americans Use Information
Overall, the findings from the survey indicate that the decision-making process from the use of AI tools has transitioned from a fledgling technology to a common part of everyday life in a surprisingly brief timeframe. That close to 80 percent of American adults are making decisions based on AI-driven information, with or without fact-checking by the user, constitutes a significant transition in Americans’ processing of online information sources.
About the Survey
The survey was conducted by Clear Spark Digital in April 2026 via research website Prolific.com. Total respondents included 2,127 adults residing in the U.S., age 18 and older. The demographic data of the sample was obtained via Prolific’s preselected screening information, which includes sex, age, education, employment, industry, and work function.
Questions about verification habits were asked only to the 2,076 respondents who indicated that they use AI tools. Of the respondents, 74% work full-time, and 20% work part-time. According to the Clear Spark Digital report, results should be interpreted as representative of a cross-section of the American adult population with some minimum amount of internet access, rather than a probability sample of the entire American adult population.
Source: FG Newswire