According to a recent survey, 75% of Americans say they only “sometimes” or “hardly ever” trust information generated by artificial intelligence (AI).
Quinnipiac University released a poll on Wednesday (April 16th), where 51% of respondents said they “sometimes” trust information generated by AI, while 24% said they “hardly ever” trust such information.
The survey by Quinnipiac University was conducted from April 3rd to 7th, with 1,562 adults surveyed, and with a margin of error of 2.5%.
The groups with the highest trust in information generated by AI were men, the younger generation, and wealthier Americans.
Among men, 7% said they “almost always” trust information generated by AI, while 21% said they trust it “most of the time”.
On the other hand, only 1% of women said they “almost always” trust artificial intelligence, with 14% trusting it “most of the time”.
There were significant differences in trust levels among respondents of different generations.
30% of Generation Z (those born between 1997 and 2012) said they “almost always” or “most of the time” trust information generated by artificial intelligence.
For Millennials (born between 1981 and 1996), Generation X (born between 1965 and 1980), Baby Boomers (born between 1946 and 1964), and the Silent Generation (born between 1928 and 1945), these numbers were 28%, 22%, 12%, and 9%, respectively.
Moreover, trust levels varied among different income groups.
Data showed that among those with a household annual income exceeding $200,000, 38% said they “almost always” or “most of the time” trust AI.
For lower-income groups, trust levels decreased. Among those who said they “almost always” or “most of the time” trust AI, 20% had an annual income below $50,000, 19% had an income of $50,000 to $100,000, and 21% had an income of $100,000 to $200,000.
When it comes to daily life, a majority of Americans (44%) believe that the disadvantages of AI outweigh the benefits, while 38% believe the benefits outweigh the disadvantages, and 18% had no opinion.
Poll experts indicated that people’s skeptical attitude is reassuring.
Brian O’Neill, an associate professor of computer science at Quinnipiac University, stated in a release, “The vast majority of Americans only trust AI-generated information sometimes or hardly ever, which is reassuring because it indicates a healthy level of skepticism when using AI as a research tool.”