Only 16% of Americans See AI as Positive for Society, Pew Poll Finds Amid Skepticism

A new Pew Research Center poll finds that just 16% of American adults believe artificial intelligence will have a positive impact on society, while 40% say it will make things worse. The findings come as usage of AI chatbots has surged, with half of U.S. adults now using them, up from 33% in the summer of 2024.

The survey, conducted among 5,119 U.S. adults and released Wednesday, shows that 63% of respondents think AI is advancing too quickly. OpenAI’s ChatGPT remains the most popular chatbot, used by 44% of Americans, up from 34% in 2024. Google’s Gemini follows at 24%, with Microsoft Copilot at 17%, Meta AI at 14%, xAI’s Grok at 8%, Anthropic’s Claude at 6%, and Character.ai at 3%.

Younger adults most skeptical despite highest usage

Usage of AI chatbots varies sharply by age. Among 18- to 29-year-olds, 66% use them, compared with 23% of those aged 65 and older. Yet younger adults are the most pessimistic: 48% of 18- to 29-year-olds think AI will negatively affect society over the next 20 years, and 37% expect a negative personal impact. In contrast, among those 50 and older, only 37% see a negative societal effect and 28% a negative personal one.

Income correlates closely with adoption. Sixty-six percent of upper-income Americans use AI chatbots, versus 49% of middle-income and 41% of lower-income adults. Asian Americans report the highest usage at 70%, while White Americans have the lowest at 46%.

• 42% of users rely on AI for information search, making it the most common use case.
• 38% of employed adults use AI at work, raising concerns about proprietary data sharing.
• 24% use AI for image or video creation, and 20% seek medical or diet advice.
• 10% turn to AI for emotional support, and 4% use it for companionship.

Trust in regulation and security remains low

The poll reveals deep distrust in oversight. Sixty-seven percent of American adults say they have little or no confidence in elected leaders to regulate AI, and 59% lack confidence in companies developing and deploying it responsibly. On security, 71% believe AI will make their personal information less secure, while only 3% expect improvement. Among those who use AI, 30% say it helps their productivity, but just 8% believe it improves their happiness.

Political divides are minor: majorities of both Democrats and Republicans express limited confidence in government or corporate stewardship of AI. With usage climbing but public sentiment souring, technology firms and policymakers face growing pressure to demonstrate responsible management of AI systems.