(TNND) — A new Pew Research Center analysis set out to understand how Americans are encountering artificial intelligence as they browse the web.

The Pew Research Center examined about 2.5 million web page visits from a group of 900 people who agreed to share their browsing data for a full month.

Most folks hit on at least some AI-related content, but much of it touched on AI at just a glancing or superficial level.

Many people encountered AI-generated summaries in their web searches.

There were limitations to the tracking. Not all web-connected devices that people used were tracked. And the Pew Research Center wasn’t able to analyze real-time algorithmically curated content on social media pages or pages that require a login to access.

Nearly everyone, over 90%, visited at least one page that mentioned AI, even if the mention was brief or unrelated to the main focus of the page, according to the Pew Research Center.

Close to half of the people visited a page where AI was the main focus of the content, but those visits weren’t frequent.

Those, for example, could be news articles that looked closely at AI-related issues.

Close to 60% of people viewed a search page, such as Google, with an AI-generated summary of results.

Just 13% of people visited the website of a generative AI tool.

And just 10% of people search for an AI-related term.

“It just confirms that AI in one fashion or another has entered the mainstream,” AI expert Anton Dahbura said of the Pew Research Center study.

Dahbura, the co-director of the Johns Hopkins Institute for Assured Autonomy, said AI is ingrained in our online world.

“And if it's not, it will be soon,” he said, noting that companies are rushing to incorporate AI both in customer service and behind the scenes.

“There's definitely fear of missing out and a perception that if I don't have AI involved somewhere ... I'll become irrelevant, or my products or services will become irrelevant,” he said.

Dahbura said the average person might not realize how much AI-generated or AI-influenced content they’re consuming online.

Daniel Schiff, a policy scientist and the co-director of the Governance and Responsible AI Lab at Purdue University, mentioned AI literacy as a concern.

People might struggle to separate reliable and unbiased information from AI-generated content that can sound authoritative but can be hit or miss.

“It shapes the information people find. It shapes what they learn. It shapes what they think,” Schiff said. “So, it's incredibly powerful for our entire kind of information synthesis and learning process.”

Dahbura expressed concern over AI-generated summaries, like those found on a Google search page or those that try and present a shopper consensus on a product page.

Say you’re shopping online for a lawn mower and you get AI-generated recommendations. How do you know if they’re reliable?

“Can you even game it by having more comments about a certain brand? And then the AI thinks that that's the consensus, that that's the majority opinion,” Dahbura said.

Dahbura said AI can be a good shortcut for people online, but you also have to be careful.

AI can hallucinate. It can reference sources that don’t exist. And it can leave out important opinions in an effort to offer a digestible consensus.

“It's not fairy dust that you can just sprinkle on things and expect it to work all the time,” he said.