Americans aren’t sold on AI, and Pew has the numbers to prove it

Wall Street can’t get enough of AI. Every earnings call is a love letter to large language models, and the market cap of companies like Nvidia and Microsoft keeps climbing. But walk into a diner in Ohio or a coffee shop in rural Oregon, and you’ll hear a very different tune.

Pew Research just dropped a new study that puts a number on that disconnect. Only 16 percent of Americans think AI will have a positive impact on society. That’s not a typo. Sixteen percent. The rest are either neutral, uncertain, or outright negative.

I’ve been watching Pew’s tech sentiment data for years, and this is one of the lowest numbers I’ve seen for any emerging technology at this stage of adoption. For context, when Pew asked similar questions about the internet in the late 90s, optimism was far higher. AI has a PR problem, and it’s not subtle.

The study doesn’t break down exactly why people feel this way, but you don’t need a PhD in sociology to guess. Job displacement fears are real. The endless parade of AI-generated spam, deepfake scams, and creepy customer service chatbots isn’t helping. And the tech industry’s habit of promising utopia while delivering chaos has worn thin.

What’s interesting is that the pessimism cuts across demographics. It’s not just older folks who don’t “get it.” Younger Americans, who grew up with algorithms shaping their social media feeds, are also skeptical. Pew’s data shows that even among those who use AI tools regularly, the positive sentiment barely cracks 30 percent.

There’s also a trust issue that’s hard to ignore. The same companies pushing AI are the ones that spent the last decade mishandling our data, amplifying misinformation, and designing addictive products. Why would anyone believe they’ll get this one right?

I’m not saying AI is doomed. The technology itself has real potential, especially in niche areas like medical imaging, logistics optimization, and scientific research. But the mainstream perception is being shaped by the worst use cases, not the best ones. And the industry’s relentless hype cycle is making things worse.

If you’re building or investing in AI, this data should worry you. It’s not enough to make better models. You need to make people believe those models will actually improve their lives. Right now, most Americans aren’t buying what you’re selling.