AI runs on the open internet. So does consumers’ ultimate trust.

NEW YORK— Generative AI may still feel like a relatively new way to discover information, but EMarketer expects it to become an increasingly important part of how consumers shop—and how brands are found. 

Opening EMarketer’s standing-room-only Future of Digital Summit, VP of Insights Stephanie Paterik shared new data from their latest report showing that one in three U.S. consumers will use generative AI to shop by 2030. Those consumers skew toward households earning $200,000 or more, parents with children at home, women, urban consumers and people in the Northeast, she said.  

But the rise of AI doesn’t mean consumers are abandoning the rest of the internet. 

In fact, the emerging behavior suggests something more complicated: AI may become another layer in the discovery journey, while websites, publishers and the broader open internet continue to provide the evidence consumers and AI systems use to make decisions. 
 
At least four sessions during the day-long event built the case itself almost by accident: the more generative AI reshapes how people shop and search, the more it depends on the open web to do it. Not as a relic being phased out, but as the raw material AI tools have no substitute for, and as the thing consumers still trust more than the AI tools built on top of it.

AI’s answers are made of open web content

Paterik described it as an expanding “ecosystem of signals.” Three sources that generative AI is leaning on most heavily, she said, include brand websites, publishers (including news), and industry sources. 

That creates a new challenge for marketers: It isn’t enough for a brand to control what it says about itself. Increasingly, brands need other credible sources to reinforce those messages. 
 
“We’re going to see the rise of the trust economy,” Paterik says. “Consumers still double-check AI’s answers. Brand equity still matters. The brand wars happening outside of AI, still matters.” 

EMarketer principal analyst Nate Elliott offered another reason not to write off the traditional web.  

AI currently accounts for only a small share of search activity. In fact, EMarketer data points to AI accounting for 2.9% of search currently. Elliott joked that “the majority of people will spend more time using the bathroom today than using AI,” which got an audience chuckle. But he also predicted that the balance will change, saying that it’s “growing faster than any channel we’ve tracked so far.”  

So, while search may be slowly condensing the consumer journey and collapsing the traditional shopping funnel, the definition of search — and the places where consumers encounter information about brands — is expanding. 

That distinction is important when considering new research from The Trade Desk Intelligence. Consumers are 1.3 times more likely to fully trust the open internet than AI tools and 2.1 times more likely to use the open internet when they make a big decision, according to new research from The Trade Desk Intelligence. Nearly 9 in 10 consumers surveyed say they are not fully comfortable relying solely on AI tools for big decisions.  
 
Consumers are right to verify. On a separate panel, Alex Sherman, CEO of Bluefish, which works with brands to get them visible within LLM search results, confirmed that they see answers from searches containing a large number of inaccuracies, roughly 10–15% of answers. This, in turn, is spiking the desire for the models themselves to find that quality content, he says.

The website stops being the whole pitch 

 The clearest statement of the day’s throughline came from Heather Physioc, Chief Discovery Officer at VML. It was that a brand’s own website was never the whole story, and AI has made that harder to ignore. 

“The website alone is not sufficient to perform,” Physioc said. “It’s an ecosystem of evidence where earned media validates what you’re saying. Someone on the outside of the brand almost has to be saying the same thing. All of these channels are coming together with that connective tissue.” 

That “ecosystem of evidence” is the open web. A brand’s own site can say anything it wants about itself. What AI-driven discovery increasingly weighs is whether the rest of the open web corroborates it, which is exactly what research from Melissa Hutter, senior director of global market research and insights at The Trade Desk, found consumers doing manually — especially once the stakes get high enough. 

Hutter says brands should invest more in the premium content that makes up the open internet for this reason. Of those surveyed in the TTD Intelligence study, two-thirds of consumers said AI search tools gave them the surface level facts, reviews, articles and opinions around the information they were searching for. The TTD Intelligence study also found that advertising on premium media resulted in a 20% lift in advertiser trust, 37% lift in advertiser quality and 41% lift in advertiser purchase intent.  

“AI will make your brand known, but premium media will build affinity and choice,” said Hutter. 
 
“You need to think of the sum total of where your brand is appearing,” noted Caroline Giegerich, VP of AI and Marketing Innovation at the IAB, on a different panel. 

Looking across collective channels is an AI learning across the board. On the panel Engineering the Modern Marketing Org, Shivani Srivastava, VP of Marketing Transformation at Spectrum, has started treating search, social, brand, and product as one connected system rather than optimizing any single one alone. A brand can’t outsource its whole presence to one AI interface any more than to one platform. 

The picture is consistent: AI is growing its share of discovery by interpreting open web content, not replacing the need for it. When the decision actually matters, consumers still go around the AI layer to the source material. The open web isn’t losing ground to AI. It’s the thing AI has to keep citing, and the thing consumers still reach for directly to trust what AI tells them in the first place.