Only 16% of Americans Think AI Will Be Positive for Society

There’s a striking contradiction sitting at the heart of America’s relationship with artificial intelligence right now. Millions of people are using it every single day — and a growing majority of them think it’s going to make things worse. A sweeping new study from Pew Research Center puts a number on that unease: only 16% of Americans believe the AI impact on society will be a positive one over the next two decades. That’s not a rounding error. That’s a full-blown trust deficit.

A Nation Using AI It Doesn’t Trust

The Pew findings are striking precisely because they arrive during what could fairly be called peak AI hype. IPO filings, billion-dollar funding rounds, product launches every other week — Silicon Valley has never been louder about AI’s transformative potential. And yet, out in the broader country, something very different is happening. Roughly 40% of Americans expect the AI impact on society to be net negative, while most of the rest are sitting somewhere in the middle — not optimistic, not quite alarmed, just deeply unconvinced.

This isn’t apathy, either. Nearly two-thirds of Americans believe AI is developing too quickly. That’s a specific, considered concern — not the vague technophobia that the industry sometimes dismisses public skepticism as. People are paying attention. They just don’t like what they’re seeing. When researchers ask Americans to describe the AI impact on society in their own words, the answers tend to cluster around job displacement, misinformation, and a loss of human control — not abstract fears, but concrete ones.

Who Trusts AI Least — and It’s Not Who You’d Expect

The conventional narrative assumes younger people are AI’s biggest cheerleaders. The Pew data blows that assumption apart. Adults under 30 are actually the most pessimistic cohort in the entire study, with just 14% believing AI will have a positive AI impact on society. That’s even lower than the already grim national average. These are people who’ve grown up with algorithmic feeds, content moderation failures, and the slow erosion of online spaces they used to enjoy. They may be the heaviest users of AI tools, but they’re not its most faithful believers.

There’s also a clear gender divide worth paying attention to. Men are more likely to use AI chatbots daily — 27% versus 20% for women — and they skew toward broader platform adoption, with higher usage rates for tools like Microsoft’s Copilot and Elon Musk’s Grok. Women, by contrast, report more skepticism overall about the AI impact on society. That’s not a trivial gap. It likely reflects a broader pattern in which the people building and promoting these tools look very different from the broader population being asked to adopt them.

The Regulation Gap Nobody Wants to Talk About

Here’s where the AI impact on society concern really crystallizes. A full 67% of Americans don’t believe the US government will do anything meaningful to regulate AI. And 59% don’t trust companies to develop it responsibly on their own either. Think about that for a moment — a majority of the country has concluded that neither the public nor private sector is up to the job of keeping this technology in check. That’s not a gap in communication or a PR problem. That’s a governance crisis in slow motion.

Washington has made some noise. The Biden administration’s executive order on AI and various Senate hearings with tech CEOs generated headlines, but they haven’t translated into the kind of concrete legislative framework that, say, the EU has been building with its AI Act. Americans appear to have noticed. Skepticism about government action isn’t irrational — it’s a pattern recognition exercise, and the pattern isn’t encouraging. Without meaningful oversight, public concern about the AI impact on society is unlikely to ease.

ChatGPT’s Dominance — and What People Are Actually Doing With It

Despite all the pessimism about AI impact on society, usage is climbing steadily. Around 25% of Americans now say they use AI chatbots on a daily basis, and the tool they’re reaching for most is overwhelmingly ChatGPT. Pew reports that 44% of US adults now use OpenAI’s chatbot — more than double the figure from 2023. That growth curve is genuinely remarkable for a product that didn’t exist in the public consciousness three years ago.

The runner-up list reveals a lot about the competitive landscape. Google’s Gemini sits at 24%, which is respectable but arguably disappointing for a company with Google’s distribution advantages. Microsoft’s Copilot follows at 17%, Meta AI at 14%, and then a sharp drop-off to Grok at 8%, Anthropic’s Claude at 6%, and Character.ai at 3%. Claude’s relatively low consumer footprint is interesting — Anthropic has attracted enormous enterprise interest and serious press coverage, but it hasn’t cracked mainstream adoption the way OpenAI has.

Most people using these tools are doing so for research or work-related tasks. That’s a somewhat more pragmatic use case than the breathless demos of AI writing novels or generating marketing campaigns might suggest. People are mostly using chatbots the way they once used Google: to find things out faster. Six in ten survey respondents also told Pew they regularly read AI-generated summaries online — which, if you’ve used Google lately, tracks completely. How these everyday habits shape the broader AI impact on society over the long run is a question researchers are only beginning to answer.

Half the Country Still Isn’t Using AI at All

The other side of this story is just as important. About half of Americans say they don’t use AI in their daily lives. The non-users skew older — nearly 75% of Americans aged 65 and above say they never use AI chatbots — but the data also shows that non-users across all ages have no particular plans to start. They’re not waiting for a better product or a more intuitive interface. They’re simply not interested, and they’ve said so explicitly to Pew’s researchers.

That’s a meaningful counterweight to the idea that AI adoption is an inevitable, frictionless process. It isn’t. Roughly half the country is watching this technology roll out and choosing to stand aside. The AI impact on society will look very different depending on whether that half eventually gets pulled in — or continues to opt out entirely. Whether that changes as AI becomes more deeply embedded in products people already use — search engines, productivity software, social platforms — remains to be seen. But the assumption that skeptics will eventually come around hasn’t been earned.

What This Actually Means for the AI Industry

The broader AI impact on society debate won’t be settled by another product launch or a more impressive benchmark. The Pew study points to something deeper: a fundamental mismatch between what the industry is building and what most Americans actually want from it. You can’t close that gap with better marketing. The concerns about speed, safety, and oversight are real, and they’re coming from people who are already using these tools — not from those who’ve never tried them.

AI companies have a choice. They can continue to define success by model capability and market cap, or they can start engaging seriously with the question of what a technology that the majority of its own users views with suspicion actually owes to society. The AI impact on society is ultimately determined not just by what these systems can do, but by who governs them and in whose interests. The trust deficit revealed by Pew isn’t a communications failure. It’s a product roadmap problem — and until the industry treats it that way, that 16% optimism figure is unlikely to move much.

Source: TechCrunch

Frequently Asked Questions