- 14 hours ago
Category
🗞
NewsTranscript
00:00Debbie, I'm excited because you know AI, you understand AI, you've worked in AI, and now you're going to explain AI exactly how you think AI will go in the future.
00:09First of all, I mean, if you're an investor, is there too much valuation in AI? Are they priced to perfection?
00:15Or do you think really it will transform the world more than we think now?
00:20Yeah, I think we are in such an incredibly exciting moment with AI.
00:24It is really the most profound technology of our lifetimes, and I think it has the potential to help us rethink literally everything.
00:32The potential for economic impact, societal benefit is honestly boundless.
00:38When we think about the economic impact for Europe in particular, it's around a 1.2 trillion euro estimate of the productivity growth that can be unlocked for the region,
00:47which I think is urgently needed given the growth trends that we've seen in the region kind of over the last 10 years.
00:53And the difference of relative growth rates, certainly between here and the rest of the world.
00:57And then there's the societal benefits, which you think about health, you think about sustainability, and the opportunity, I think, to really rethink what's possible there.
01:05We have AlphaFold, which is the world's first AI model that won a Nobel Prize.
01:10It's basically helping to accelerate drug discovery by billions of years.
01:14And so these kind of benefits, I think, mean that this platform shift that we are living through, I think, is going to create untold amounts of value for the world,
01:23but also for the businesses and for economies, which is why I'm incredibly optimistic in this moment for Europe, but also for the whole world.
01:29I mean, when you talk about productivity, I know people keep on saying, I'm going to be 20% more productive.
01:34Is it across industries?
01:36It is.
01:37It is.
01:37I mean, one of the things about AI is it is a general-purpose technology.
01:41So like electricity or these kind of transformational technologies, they can be applied to so many different places.
01:47And if you think about the way people are using it in their own lives, many people are probably using everything from within our portfolio, you know, search and YouTube.
01:55You might be interacting with Gemini.
01:56You might be a Google Cloud customer.
01:58You might be building foundational solutions as a developer.
02:02However, all of those different use cases are enabled by AI.
02:05And within businesses, we see everything from marketing, customer service, in this industry, fraud detection.
02:11And it has so many different applications that it is, and that is one of the reasons why it's a transformative technology,
02:17because it can be applied to so many different spaces.
02:20How do you use it?
02:21So, you know, every day.
02:22I mean, I use it all the time, every day.
02:24And I think about, there are the subtle ways, you know, when people say, oh, I don't really use AI.
02:28And if you use Google Maps, if anyone used Google Maps to get here and you followed the little green dot and it told you either the most efficient route, the most fuel efficient route,
02:37I mean, all of those things are essentially using AI because they're taking tons of data and they're helping you understand based on traffic patterns, based on mobile usage, et cetera.
02:44So those are small ways everyone's using AI, and I'm certainly using it.
02:49In terms of the kind of core productivity use cases, I use it literally all day long.
02:54I'm using it to research topics that I don't know that much about.
02:57I'm using it to summarize a meeting I was in, draft a thank you note based on the notes from the meeting.
03:02I'm using it to prepare performance evaluations based on tons of notes and data that I have.
03:08I don't know if anyone's used Notebook LM, but Notebook LM allows people to basically become an instant, the tool becomes an instant expert in your documents.
03:16And you can then listen to a podcast about them, create mind maps about them.
03:20So it's perfect for education use cases, learning new things.
03:24So, I mean, there's a myriad ways that I personally am using it.
03:26I mean, we're in this race, and Google's right in the middle of it, right, to try and be the thing that everybody uses.
03:33Are we just at the start of the race?
03:35Are we, you know, how do you know the winners and losers right now?
03:38Because you must think about this day in, day out.
03:40I am incredibly optimistic about Google's position right now.
03:44I think that one of the things that's very unique about Google is that we are really end-to-end integrated.
03:48So we build the chips, the TPUs that actually power AI.
03:53We build data centers and run them.
03:56We have the world's leading frontier models with Google DeepMind just at King's Cross, led by Demis Esavis, Sir Demis Esavis.
04:04We then have 15 different products that each reach half a billion users.
04:08So whether it's YouTube, search, maps, photos.
04:11So I feel really good about our position in that space.
04:13And we have ways that enterprises can work with us, so Google Cloud, for example, or our ad solutions.
04:18So, you know, obviously it's an incredibly exciting, invigorating moment in the space.
04:23And the pace of innovation, I think, has never been more thrilling.
04:28But I think that Google has all the ingredients to succeed.
04:31How do you think about job losses or safety for some of the AI tools?
04:34You know, I hear the anxiety, and I understand the anxiety about jobs.
04:39I think that we are certainly going to live through a massive transition about how people learn to work alongside these tools.
04:46I mean, the reality is I do think there will be some number of jobs that are displaced.
04:50I think it is a smaller number maybe than some anticipate.
04:53We estimate it's around 7%.
04:55And I think what's important is that we have programs that help those 7% transition into new fields of work.
05:02I think what is more likely is that most of us are going to learn to work alongside these tools.
05:07So we estimate about two-thirds of us.
05:08If you think about the tasks you do in your daily life, pretty much every industry has tasks that can be enhanced by AI.
05:14So I often think about some of the most manual kind of work areas, like a construction site.
05:21And people think, well, how is AI going to transform construction?
05:23Well, think about scheduling people to actually show up at a building site or doing payments and the back office operations.
05:30Basically, when you look at every industry, there are administrative tasks that are happening everywhere,
05:34and that's what AI is most helpful in improving.
05:39And then I think there's this whole category of jobs that haven't ever been created yet.
05:42So I think about, I graduated from university about 30 years ago, and I think about the kinds of jobs that existed.
05:48And now if you go on LinkedIn, I mean, there were no such thing as a web developer.
05:52Now there's front-end developers, back-end developers.
05:54You can do online marketing, email marketing, prompt engine.
05:56I mean, there's this whole new field.
05:58And I think what we don't know yet is what are those new jobs of the future that are going to get created.
06:03What would you study today if you had to go back, given what you know now?
06:06I mean, I would study what I'm passionate about.
06:08I mean, ironically, I studied international relations and economics, which, given the fact that I'm...
06:12Perfect bloomberg.
06:12Exactly.
06:13I feel like I'm in the perfect place.
06:15I started my career in investment banking, actually, in financial services.
06:18So, you know, I really do.
06:20I have a 17-year-old, and I talk to her a lot about what is she passionate about, what she's interested in.
06:25And she's actually very interested in the combination of sociology and data science, which I think is kind of perfect, right?
06:30Understanding people, understanding how you can use data to understand people better.
06:33But I really do think it's about what people are passionate about is a good starting place.
06:37What's your take on European regulation for AI?
06:40So there's, again, quite a lot of anxiousness about how to deal, especially with teenagers and some of the things that could go badly.
06:46Yeah.
06:46Look, I would say my number one worry in the role that I'm in, which is leading Google across Europe, Middle East, and Africa, is the degree to which we can bring the best of our products, the best of these capabilities, and unlock this potential for Europe in particular.
07:01I think that we need to make sure regulations are in place, because I think this technology is incredibly powerful, and it deserves to be regulated, but it needs to be regulated well.
07:10And I think currently in Europe, there's something like 100 different pieces of digital legislation that have been written since 2019.
07:17They conflict with each other.
07:18They are not clear on what the outcome is.
07:20And I think there's a real need for simplification and harmonization so that we understand how do we actually get our best products.
07:27When I think about, hopefully, many of you, if you're based in the UK, you've probably been interacting with tools like AI Overviews, which is kind of the way that we give you a good summary starting place for a search query, and that allow you to dive deeper.
07:39Products like that, they came nine months later to Europe, because we had to make sure we were getting it right for what the expectations were around the DMA, the DSA, et cetera.
07:49And in France, they are still not launched.
07:51And I think that's problematic.
07:53And you can think about that and how that applies to tools for businesses.
07:56And can small businesses do the things that businesses in other parts of the world are benefiting from?
08:01And I really think it's problematic that we have so much regulation that is in conflict with one another.
08:06So I am eager to engage in my new role with sort of, I would say, industry broadly, because whenever I talk to business leaders, there's universal frustration, I would say.
08:17About 60%, I think, of a recent survey of business leaders say they'd like simplification of regulation in the European Union.
08:25I'd like to join arms with other fellow interested parties and really advocate for a simpler regime here.
08:30Debbie, is there a danger that European chief executives then also spend less in AI and technology for their firms and therefore get left behind?
08:38Yes, that is the worry.
08:41Do you speak that a lot with chief executives?
08:45I do.
08:46And it's less that I worry that the chief executives are not understanding the opportunity.
08:50It's more the context in which they operate and whether or not they have the tools available to them that other business leaders in other parts of the world have access.
08:58I would say my biggest advice to business leaders, it's been a very interesting sort of three-year arc, I would say, in these conversations.
09:05It started with, wow, AI is super cool and I need an AI strategy.
09:10And then people started to realize, actually, no, what you need is to understand how AI can supercharge your strategy.
09:15And it's not that you need something separate and different and aside.
09:18It's, as I had described, it's a general purpose technology and it can be applied to lots of different use cases.
09:25And as a leader, the responsibility is to understand what are the value creation mechanisms of my business?
09:30How do I grow?
09:31How do I drive more productivity?
09:33What are the biggest areas to get after?
09:34And I think now is the time for people to be in action.
09:38And the evolution I've seen is sort of from, like, let's test a lot of different things.
09:44And I think now we're in a phase where people need to be testing, learning, and then scaling what's working.
09:48A lot of the conversation is around, help me with the defined use cases that you're finding success in other places.
09:54And actually, for the financial services industry, just yesterday, this research from Google Cloud really shared some very specific examples where in financial services we're seeing the best results, which is kind of these use cases around customer service, which is a proving use case, I would say, across industries.
10:11It's in marketing, marketing, whether it's about segmenting audiences, if it's about making creative more quickly.
10:18It's about basically getting out and understanding what's delivering value and doubling down on that more quickly.
10:23It's also around fraud, which is a very financial services-specific use case.
10:27But understanding patterns of payments and things like that, that's a great use case.
10:31We've seen billions of dollars really being created for financial services firms.
10:35And then it's in a lot of the administrative functions, so whether it's accounting or finance, these parts of the business that basically can use AI tools to enhance their workflows.
10:42I mean, what I hear a lot from financial chief executives is, but, you know, what if I don't have the junior analyst understanding what I went through to learn the business?
10:50So how do you respond to, I don't know if it's a criticism, but to those concerns?
10:54Yeah, look, I think there's a real question about, and again, as I was talking about my daughter, you know, I think about how do people get started in these entry-level jobs?
11:03And I think what we're seeing, for example, at Google is we're still hiring entry-level engineers at pace, and 30% of our code base is now initiated by AI, reviewed by humans.
11:15But the kinds of work that we're having those early-entry employees do is actually much higher-value work than some of the bug fixes and things that typically you would have had your starting engineers working on.
11:27So, you know, I do think that these entry roles will evolve and change, and I think they should, honestly.
11:33But I think the opportunity is to think about how do we reinvent what our young people are working on as they start out in the workforce.
11:39I know you started by talking about health care.
11:41Now, I'm very familiar with DeepMind, a lot of people are, some aren't.
11:45Is it, do people understand, actually, how, you know, what this means for research?
11:49Yeah.
11:49I mean, for example, protein folding.
11:51Yes, yes.
11:52I mean, I'll give the quick explanation of alpha folds, because for me, when I was learning about it, I had to go back to my high school biology and remember what exactly is a protein and how does that relate to anything.
12:02But proteins are basically the fundamental building blocks of life.
12:05And the way I often think about it is, we remember the COVID, sorry to bring everyone back there, but we saw those images of the spiky protein, and everyone was talking about how do we create a vaccine that can attach to those spikes.
12:15That's basically a protein.
12:17And so you imagine if you're in drug development, you have to figure out how do you develop something that attaches to spikes.
12:22What the team at DeepMind was able to do is, at the time, there were about 150,000 known protein structures, and if you were in a biology lab or a chemistry lab, you would basically understand those 150,000 and try to build drug compounds that attach to those proteins.
12:37What they basically did is start with this corpus of 150,000 and basically model all the potential proteins that could exist in the world.
12:45And it's something like 200 million different kinds of proteins.
12:47And so what that's enabled is now all these edge case kind of disease types, or things like drought-resistant crop development in Africa, or malarial-resistant kind of cells, and how you actually could develop drugs or solutions in the food space that would actually address those needs.
13:08And you now have millions, I think it's something like 3 million different labs around the world that are actually using AlphaFold.
13:14And if you talk to anyone who works in biology or chemistry, they are all using these in everyday work to accelerate the ability to develop drugs.
13:23So it's pretty inspiring work that's happening.
13:26And I think, honestly, what's so exciting about AI is it's transforming on so many different levels.
13:31Healthcare, sustainability, economic outcomes, personal productivity, joy.
13:37I hope everyone's had a chance to play around with Nano Banana.
13:39Nano Banana, anyone?
13:41This is the...
13:42It gave me fake lashes.
13:43We tried it in the green room.
13:45I mean, it's fun.
13:46It's super fun.
13:47So it's kind of the range of experiences people can have with AI.
13:50It's very exciting.
13:50Debbie, I mean, on the other side, how do you think of infrastructure?
13:53Because we're using so much power.
13:54Yes.
13:55And if you want to go through the transition and be great, you need to build sustainable infrastructure.
14:00Is that like Google's job, or is it public policy?
14:03I mean, we are working alongside kind of the entire ecosystem to make sure there's enough energy to power this transition, and in a green way.
14:11So whether it's nuclear or wind or solar, we are heavily invested in all of those areas.
14:17It's another area where I feel like I want to see European governments actually leading the way here and helping us find the path forward.
14:25We're working a lot, and we're based here in the UK today, so a lot of our work with the UK government is trying to identify where are those sites that we can harness the incredible wind power,
14:34but then also get it attached to the grid quickly so that it can actually power the data centers that are needed to bring all of these good features to the world.
Recommended
5:32
|
Up next
19:43
1:58
1:57
1:46
4:35
4:14
6:14
5:22
2:42
5:20
4:28
1:57
19:23
4:25
6:01
14:25
5:34
5:19
4:46
2:34
Comments
1