Google exec says link lists aren’t going away as search transforms

In recent months, AI has fundamentally changed what it means to google something—and the transformation isn’t done.

The search engine made a slew of announcements at its developer conference earlier this year that included the first change to the search box in a quarter century. The upshot is a more AI-centric experience with options for news alert-style information agents that automatically scour the web for you.

Though Google has clarified that the familiar list of links isn’t going away, multiple headlines have called it the end of search as we know it. But Reid insisted that Google still cares about the health of the internet as a whole.

“Search is not a useful product if the web is not a thriving product, and so I think we probably care about that more than any other company in the world,” Google VP of Search Liz Reid told us.

We spoke with Reid about what this means for how you browse the web, what it means for websites that depend on Google for survival, and how it will change the ads you see.

This interview has been edited for length and clarity.

Do you think that there’s a time in the future, amid all this AI change, that the familiar list of links could go away entirely? Or that there could be a different format for funneling users to the web?

I don’t see any time in which the links go away. I think what we actually see is that people don’t want AI or the web; they want both. There’s a lot of topics where AI really gets you started, and then people want to engage more. They value the expertise, the opinions, the taste of how to think about a particular topic. This is a time in which AI is growing, and it’s a time in which newsletters and podcasts and videos and all these new formats seem to also be blossoming because people really want to hear that expertise. So we want to continue to make that thrive.

We have for a long time continued to innovate on how you can surface links on the web. So we have the blue links, but we have, at times, we have the top stories part. We have what people are saying in AI Mode and in AI Overviews.

If we know users want to connect to that information, how do we make it as easy as possible? How do we make it as seamless as possible? And we’ll continue to innovate and we announced some updates pretty recently on how we’re showcasing links in additional ways. But I think that is kind of what makes search at its heart—not just knowing the information about the web, but connecting people to content on the web. And I don’t see that going away.

Obviously, online publishers are seeing click-through rates decline with these changes, and some publishers are even considering blocking all Google crawlers. In terms of making sure that the open web is still a healthy place to supply these AI tools and search, how do you think about that in the long term?

We do spend a lot of time thinking about this. We try to really design our AI experiences in sort of a web-forward way. We have teams whose job is to think about: What are great links to surface more effectively? How do you show them in a way that’s great for folks? We’ve innovated; one example is in AI Overviews, we’ll now often show on some timely stories, a big carousel of publishers that we didn’t have three months ago. So we’re constantly trying to figure out how we innovate on that.

The good news is that people still want to connect with other people’s content, and so we can facilitate that. And as long as people make great content that other people want to read, then we believe we can continue to foster that.

Companies like Microsoft are experimenting with publisher marketplaces or other compensation schemes. Is that something you’re considering?

We have a blog post in which we’ve talked about some of the various news deals and other things we have.

We have a variety of different ways—we have a long history of deals with news and others for great content. In general, we’re all still—I mean, at times it can feel like a long time—but we’re still pretty early in the space, and so continue to innovate.

In terms of the revenue side, how has this changed Google’s ad model? Do you feel like you’ve kind of cracked advertising in generative AI?

What we are seeing is as people ask longer queries, they’re specifying more of their intent. What are they really trying to do? What’s important to them? And that sort of allows new opportunities to think about ads targeting.

There’s some efforts like direct offers, in which…the advertiser can give a more specific ad for you based on what it is. And so, an example is like you’re searching for a hotel, you’re searching with a family. Maybe a hotel wants to offer free breakfast. They’re not going to give free breakfast to the person that they think is traveling for business who will just pay, or they won’t eat. But this family might pick that hotel because of it, and so Google’s system will allow [the hotel] to say, “Oh, I have a special discount just for you for breakfast,” and that really provides new value to the user and sort of allows the hotel to get additional visitors. We both have something that’s working, and there’s lots of room for innovation going forward.

There’s always been some level of slop on the internet, and I know Google has tools to sift through that. But as generative AI tools have created this ability to churn out more content, how do you make sure that with things like the AI Overviews, the accuracy is maintained? And have you seen the accuracy of those go up overall?

To your point, there’s always been slop. Sometimes people are like, “Oh, there’s now slop on the web.” In my experience, there’s been slop for 20 years. So we have a lot of experience and teams whose job is to ensure that both we’re getting rid of spam and that that rarely surfaces, and that high-quality content continues to surface. And when we do AI Overviews or AI Mode, we actually make use of those underlying search ranking systems so that it’s the highest-quality sites going into AI Overviews. Overall, the quality of the AI Overviews has continued to improve and has improved considerably since we first launched. The models are much better, for one, so that really helps. But also, we’ve just learned a lot and continue to grow on what the techniques are [and] how you work with the models to ensure that you can actually really ground the data in high-quality information.

Are Gemini and AI Mode search still different experiences, or are they converging into the same thing?

We think of them as having two different North Stars. Gemini is really focused on being this assistant, and search AI Mode is really focused on a generative AI search experience. And so for search, that means the importance of connecting to the web is vital. And so the way that we show the web is different…we tend to focus on information use cases more heavily. Versus with Gemini, we’ll focus on a lot of productivity and creativity use cases.

The use cases overlap. We have a view that you should meet users where you are, so you don’t go and say, like, “Oops, you asked the question in the wrong product. Try again.” We will try and answer across both of them the full breadth of questions that come to our products, but we will tune them differently, and really tune search to focus on optimizing for information use cases and optimizing for connecting people to the web.

Where is Google search headed in the long term, and what are some of the biggest ways we can expect it to change further?

There’s a couple of things that we see. Experiences are becoming more personal and custom for you, for each one. It’s not just that you can ask that longer question, but that we start to understand how to make information most useful for you. We see people learn differently. We have people who love videos and people who hate videos. We have people who are experts in the financial jargon and not in the medical jargon. We have people who speak different languages.

And so, with some of the things like generative UI…the job is not just for you to technically access information, but for it to be useful to you—that’s an important part of the mission. How do we present information in ways that are much more useful? That sense of building a thing that feels designed for you, and not just designed for 2 billion people, to make it easier for you to understand and grasp the information will continue.

Also, agents and how they show up in search in terms of just, how do they offload more of the work you don’t want to do? I think people have parts of searching where they enjoy and they’re delighted, and then they have parts of tasks that feel like there’s a lot of toil. And so, how can you offload those parts, and then make it easy for people to spend their time on the things they want to enjoy?

We’ll see that people will ask harder and harder questions. That’s been the evolution of search. You get better at answering the question. They’re like, “OK, I’m gonna raise the bar on you,” and so I think we’ll continue to see people go more and more to, what’s the real question they’re trying to get done, what’s the task they’re trying to get done.