- 6 hours ago
Category
🗞
NewsTranscript
00:00This is a story about anticipating your every want.
00:03Last week, we told you how AI is cutting into traffic for Internet publishers
00:07and what that could mean for their business.
00:10But Google isn't standing still for all the challenges from chatbots.
00:14Last month, it announced the biggest changes to its search approach in over 20 years.
00:19Nick Foxx is Google's Senior Vice President for Knowledge and Information
00:23and the man responsible for the changes.
00:26With your experience, rank order, which is probably unfair,
00:30this change with AI compared to other innovations you've done?
00:34Oh, I think number one for sure.
00:36For Google, that transformation includes AI-generated answers,
00:40conversational search, and new tools that can reason through complex questions,
00:45monitor information across the web, and even complete tasks on a user's behalf.
00:50How would you describe the change?
00:52You know, just to take a step back,
00:54AI is the best thing that's ever happened to search.
00:56And, you know, a lot of people are questioning that.
00:57A lot of people are saying, hey, what does AI mean for search?
01:00And what we see, you know, the vision of search
01:02is you should be able to ask whatever question is on your mind.
01:04And the way I think about that is we have probably thousands of questions
01:07that go through our head any given day.
01:10We ask maybe a few of them, maybe we ask five of them, maybe we ask ten of them.
01:13Why do we not ask them all?
01:14Well, we don't know if there will be a helpful response to it.
01:17It might be hard to ask that question.
01:18So our big idea is you should be able to ask anything.
01:21And AI really supercharges that.
01:23AI enables people to ask whatever question is on their mind and get a helpful response.
01:28And so that's what we're pushing on.
01:29The quality of the user experience is a priority for Google,
01:32even as it's changed its search engine to put AI at its core.
01:37And Rand Fishkin, who's specialized in the business of online search,
01:41says that improved experience is very much central to Google's business.
01:45Instead, they push those down and they put the instant answer from AI right up at the top.
01:51They call them AI overviews.
01:52In the long run, it's better for their user experience,
01:55and it creates an addiction to search and quick answers.
02:00If they can get people searching more and instantly answering their queries,
02:04they would come back more often.
02:05They would trust Google more, and they would click on more ads in the future.
02:10So Google is hurting themselves a little tiny bit right now,
02:13costing themselves maybe a dollar that they could charge,
02:16so that in the future, they make $5, $10, $50, $500 from you.
02:21Fox does not deny the advantages to Google if its new approach to AI-driven search
02:26gives users more reason to use its product,
02:29but sees that as a feature rather than a bug.
02:32What we're seeing with users is they're asking far more questions
02:35than they've ever asked before.
02:36They're getting the approach that we've taken,
02:39the approach that really defines what Google in search is doing with AI,
02:42is that we're able to bring the best of the web together with the best of AI.
02:45We're able to bring frontier AI capabilities built on our Gemini technology
02:50in close collaboration with DeepMind,
02:52and then we're able to bring that together with the web.
02:54So people can get links, they can discover content across the web,
02:58but they can also get helpful responses that contextualize the web
03:02and also give contextualized responses that answer their question.
03:06But Fox insists that time spent on a search is not the right metric for success.
03:11He'd rather people spend less time on a particular question or search
03:15and be able to press further to get their answers, including from the web.
03:20For us, it doesn't have to be about time and to some extent attention,
03:23but we don't measure search by how many minutes are people spending on search.
03:28We measure search by how many queries are people doing,
03:31how many tasks are people doing, actually.
03:33If it takes two queries to do a task instead of five queries to do a task,
03:38we would prefer that.
03:39We actually don't want people to do more queries in a way where they're,
03:42you know, if they're doing more queries because they're thrashing
03:44and they're not getting a good experience,
03:45or they're spending more time because they're thrashing
03:47and not getting a good experience, that's not what we optimize for.
03:50We actually try to get people off Google as quickly as we can.
03:52It's kind of, you know, one of the things that I think has been a hallmark of Google.
03:57We do want people to be doing more things with Google.
03:59If we can help people with, you know,
04:01twice as many of their questions as we could, you know, before, that's great.
04:05But it should be they're doing more with Google because they're getting more done
04:08rather than they're spending more time to get the same set of things done.
04:12As AI takes on a greater role in helping us gain access to content on the web,
04:17there are those concerned about its use discouraging the creators of the content.
04:22Caitlin Petrie is author of the book, All the News That's Fit to Click.
04:25I think that it's really important to think about what's happening with
04:29people who make content and AI as this kind of longer story of a tussle or a conflict
04:37between tech companies and people who create media and information.
04:41And then the question becomes, if the business model for making original content,
04:47and particularly I'm concerned with journalism and art,
04:51but if the business model for making original stuff kind of goes away
04:55or becomes even more challenging and even more punishing than it already is,
05:00what do we do about that?
05:02You know, people today expect that when you come to Google, it's accurate, it's reliable.
05:06Fox says he has similar concerns that whatever Google does with AI and search
05:11not take away from the value of the content on the web for the sake of all involved.
05:16What does this mean beyond the users for the rest of the web, the rest of content generators?
05:23I mean, because Google was originally a way to organize the web, right?
05:27And a lot of content generators got linked back to their material.
05:31What does this new approach mean for them?
05:33So we remain committed to Google organizing the web, right?
05:36That's a core part of what we do.
05:38And I would say relative to really any other company, really any other organization out there,
05:44we care about the web incredibly, incredibly deeply.
05:47We're a company that started in the web.
05:49We're a company that still is deeply in the web.
05:51And so that is a bedrock commitment for us.
05:56And so that leads to how we actually build our products as well.
05:59So, for example, in search, it's not just AI.
06:02It's AI with links within the AI as well.
06:05We don't think the users need to make a choice, do I want AI or the web,
06:09but rather we can bring both of those together in a really effective way.
06:13So we have lots of links within the AI responses.
06:17It's not just because of a commitment to the web.
06:19We also think it's really useful for users, right?
06:21If you're interested in a topic, you want to go deep on a topic,
06:23you might get an AI summary, you might get an AI sort of overview to give you the context,
06:29but then you want to read it more deeply.
06:31You want to see original reporting on a topic.
06:34You might want to see, you know, firsthand perspective.
06:37You might want to hear from someone who's actually there.
06:40You actually, you want to get that.
06:41You don't just want to hear from the AI.
06:43And so we're cognizant of that and we design around that.
06:47People are searching more than ever.
06:48People are searching for topics they were never searching for before.
06:51That all creates opportunities for websites to get discovered that never would have before, right?
06:55If a user was never asking the question, the website never would have gotten traffic,
06:59never would have gotten a click.
07:00This is an expansionary moment, not just for search, but for the web overall.
07:03Accuracy and trust have long been the foundation of Google search.
07:07But in an era of AI-generated answers where users may never click through to that underlying source,
07:13ensuring the quality of that information has become more important than ever.
07:17Accuracy and trust are important when you're searching.
07:21How do you monitor that?
07:22Because there are a lot of people out there who actually have almost a goal of putting out information that's
07:27not reliable,
07:28that's not accurate.
07:29How do you monitor how you're doing?
07:31I mean, this is what Google has been doing for years and years.
07:35It's what we've been doing for decades.
07:36We understand the web.
07:38We understand the web very well.
07:39We understand the very early days of PageRank was that you could use the citations across the web to understand
07:46what's more valuable than others.
07:49We've obviously built on that tremendously over time.
07:52But we understand the web well.
07:53We understand what's accurate on the web.
07:57We understand what's reliable.
07:58We understand what's trusted across the web.
08:00We're also able to understand when we get it right, when we don't get it right,
08:04because we're able to understand sort of metrics across our user populations.
08:09So, for example, when we roll out a new AI experience within search,
08:13we know based on, you know, we run experiments, we run tests,
08:17and we can analyze those to see has this improved the experience or not.
08:21We only release changes that actually do improve the experience.
08:24Are there some categories of information that you have to be particularly careful about?
08:28I mean, I'm mindful of the fact we're going to have elections in the United States coming up in November,
08:31and you can have a lot of people trying to use AI and the web to misinform,
08:36either political rivals or foreign governments.
08:38What do you do to protect against that?
08:40We are particularly careful on those types of questions.
08:44So if an election type of question, we're particularly careful,
08:47a finance type of question, a health kind of question.
08:50The way we approach that is in a few ways.
08:52Number one, if we don't think the accuracy of the information is high enough,
08:55we won't show an AI response in that kind of a case.
08:58So that's kind of, that's the first level of it.
09:01The second is we invest incredibly heavily here,
09:03and we focus particularly heavily in categories where the stakes are particularly high.
09:09And then the third piece is we're clear with users about what is AI generated,
09:13and we indicate, you know, AI can make mistakes so that users can be cognizant of that as well.
09:20As AI reshapes how people find information online,
09:24it is also reshaping the competitive landscape.
09:26And while Google may dominate search today,
09:30the rise of AI has opened the door to a new wave of competitors.
09:34You come from a very strong position in search.
09:38I mean, you've become a verb in addition to a noun around the world.
09:43Do you have a comparative advantage against other AI companies in that you have search
09:48coming to AI rather than starting in AI?
09:51I think our advantage comes from a deep expertise and a deep focus on innovation and continual reinvention.
10:00I think one of the maybe understated advantages that Google has is we've been through many disruptions before.
10:07We went through a mobile disruption.
10:09We arose, you know, in the web disruption.
10:12There was Ajax technologies.
10:13There was voice.
10:13There's been sort of time after time disruptions.
10:17Each time it would be, is this the Google killer?
10:20And we have worked through those and we have really a playbook of how to work through those,
10:25which is lean into the innovation, innovate through the disruption with an excitement about what the technology can enable
10:36and a promise to build a better product for users through it.
10:39And so I think that's a core advantage through any of these transformations is to rely on that playbook.
10:46You raise, I think, a really interesting point.
10:48I mean, was there any hesitation anywhere at Google to say, we're doing pretty well in search.
10:54Why do we need to disrupt ourselves on this?
10:56It's sort of a form of innovator's dilemma.
10:59Was there any sense of, you know what, we're doing pretty well.
11:01Why rock the boat?
11:02Never, because we've seen it before and we understand that if you stand still,
11:08if you don't innovate for users, you'll become irrelevant.
11:12And that's so clear.
11:14We know that.
11:14We know that time and time again.
11:16And so we see a new technology come along, we jump on it.
11:19We invent the new technologies in most cases, but we see it, we jump on it.
11:24And we know that if we innovate for users, it will be expansionary.
11:29We know that our business will thrive through it.
11:31For Google, the choice was between disrupting its strong business and search or having someone else disrupt it.
11:39As AI yet again reshapes how information is discovered and consumed, the question will be, as it always is,
11:47whether this new innovation gives people what they want and need.
11:52We're seeing that users are loving it.
11:54And, you know, the best sign that we see is that queries are growing.
11:56We're suspecting them.