AI: Only 16% of Americans Say It Will Have Positive Impact

Only 16% of Americans think AI will have a positive impact on society during the next 20 years. Around 40% say it will have a negative impact, and a majority express skepticism about who will keep the technology in check: 67% do not believe the U.S. government will meaningfully regulate AI.

Pew Research’s survey captures a split between heavy use and low optimism. Many people report daily interaction with AI tools even as doubts about corporate and government stewardship persist.

How do Americans feel about AI?

Public sentiment is largely neutral to negative: only 16% of Americans expect AI’s impact on society during the next 20 years to be positive, roughly 40% expect it to be negative, and 67% do not believe the U.S. government will meaningfully regulate AI. Additionally, 59% of respondents said they do not trust companies to develop AI safely.

Young adults are the most pessimistic. People under 30 are less likely to expect positive outcomes, with only 14% of that cohort saying AI will have a positive impact. A separate majority view concerns the pace of development: nearly two-thirds of Americans think AI’s development is occurring too quickly.

Half of the country reports not using AI in their daily lives. Those who avoid AI tend to be older, and many say they are uninterested and have no plans to try chatbots in the future.

Who is using AI chatbots and how?

Chatbot use is widespread but uneven: about a quarter of Americans use AI chatbots daily, 44% of U.S. adults say they use ChatGPT (a figure that more than doubled since 2023), and usage falls across specific products—Gemini 24%, Copilot 17%, Meta AI 14%. Grok, Claude, and Character.ai register lower shares at 8%, 6%, and 3% respectively.

Men report higher daily chatbot use (27%) than women (20%), and while equal shares of men and women report using ChatGPT, men more commonly report use of other services such as Copilot and Grok. Those who use chatbots most often say they do so for research or work. Six in 10 respondents also said they routinely read AI-generated internet summaries, indicating AI is already changing how many people consume information.

Usage skewed by age is stark. Nearly 75% of Americans aged 65 or older say they never use AI chatbots, while people under 50 are more likely to report using them.

Why it matters

The split between heavy use and widespread distrust matters because people are relying on AI outputs while doubting the institutions that build and govern those systems. High rates of routine consumption (for example, six in 10 reading AI-generated internet summaries) mean AI can shape information flows even as 59% of respondents say they do not trust companies to develop it safely and 67% doubt meaningful government regulation. That combination raises risks around accuracy, accountability, and who benefits from AI’s rollout.

What to watch

Key signals to follow include shifts in chatbot market share (ChatGPT is currently used by 44% of U.S. adults) and any concrete U.S. regulatory steps—67% of respondents say they do not expect meaningful government action. Also watch the under-30 cohort: only 14% now expect a positive impact, and changes in their attitudes as adoption grows would be an early indicator of broader opinion trends.