An Insight, An Idea with Sundar Pichai | DAVOS AGENDA 2021

Sundar Pichai
1 follower
Good evening, everyone out there in the virtual realm. My name is Miriam Meckel and I'm very
delighted to host this conversation tonight with the CEO of Alphabet and Google, Sundar Pichai.
Good morning, Sundar. Hi, Miriam. How are you? I'm good, actually. I'm far away, but I'm good. It
would have been much nicer to talk to each other in person, but that'll be possible hopefully next
year again. Well, I'm glad we can do this virtually. As you said, hopefully it changes
over the course of the year. I think so. Maybe we just jump into our conversation about insights
and ideas with connecting the last year and the year that just began. A year of restrictions
lies behind us. Recession is still going on and 2021 just begun. How do you encourage the people
around you, the teams at Google, maybe even yourself, to somehow still trust in the power
of technology, in the power of technological innovation? Give us a streak of light, please.
It's something I think about a lot. I'm a technology optimist, not because I believe
in technology, but I believe in how people come together to use technology for good.
From my personal experience, it's always resonated with me because technology created
opportunities in my personal life. I see countless examples around the world of how
technology is driving progress. COVID itself has been a moment of reflection and while you see all
the hard impact around the world, you also see the role technology plays and how most of the
economy is still functioning around the world. The vaccine development, you realize it's happening on
a foundation of technological progress over the past many decades. When you see advances like MRNA
come into fruition and you know it's going to be applicable to other heart problems down the line.
Within Alphabet, as DeepMind made progress with AlphaFold, you had a moment where AI in a few
years made an advance. It's a 50-year breakthrough in terms of protein folding and I can see its
potential to streamline aspects of drug development. It gives me hope that 10 to
20 years down the line, for the next challenge, it's a foundation which will help drive
progress and that's how the virtuous cycle continues. As a company, I take a
long-term view and whenever you step back, you see the long arc of technological progress.
That's what I talk to my teams about and gives me optimism through this very difficult time.
Let's break this down a bit. I'm totally with you regarding the messenger RNA. I think that's
really a breakthrough. I mean a vaccine regularly takes like five years and longer to be developed,
to be researched and now we have it already in different forms. On the other hand, what I
observed is that artificial intelligence, for example, didn't play such a big role in
the analysis of COVID in many regards. It was kind of underwhelming, let me put it like that,
and now we have huge difficulties with the distribution of the vaccine. Couldn't AI
help with that? You know, I think today we have the technology tools to help with
cloud computing and machine learning and algorithms exist today. I think the real
potential of AI will come. We are still in very early days of AI and I think all of this will
come into play in a 10 to 20 year time frame. But having said that, I see countless. So for example,
when you look at how did we as a company deal with making sure users got the right information
around COVID and preventing misinformation, we did rely on AI algorithms to make a lot of
decisions and it's a difficult area and I think AI contributed significantly to help us make
progress there. But the pandemic is a one in a hundred year event globally and I don't think
AI played, you're right, any significant role in helping us navigate it. It is the foundations of
all the technological progress over the past many years and I think AI is laying the foundation for
us to tackle future problems or current problems like climate change and how do you solve
the next advances in healthcare and so on. If you figure a pandemic in let's say 10 or 15 years,
what would then be different with the progress of AI you were just describing for our future?
You know, I mean, you know, our ability to, you know, detect, plan, coordinate,
you know, and drive the next generation of therapeutics and vaccines, etc. Everything
will, you know, assuming as humanity we use this moment to learn from it and apply our,
you know, creativity and resources, you know, I think we will tackle it definitely
better the next time around and so I see all of this in a cross-cutting way
help drive progress and, you know, so for example, you can see even through this pandemic
we had shortage of masks. We learned what is the right mask guidance through the course of
pandemic. These are all, you know, simple things to get right the next time and, you know, and,
you know, how do you accelerate therapeutics? You know, I think AI will play a big role
by the time we have to deal with the next one. So, it's not just about machine learning systems,
it's still about human learning systems as well to tackle issues like that. Absolutely. Now,
you just mentioned the problem of misinformation around COVID and there was quite a lot.
I would like to take a moment to take it to a little bit of a meta level and also look at what
happened during the start of this year. In many ways, 2021 has started very unexpectedly
or maybe not, but maybe yes. What went through your mind when you saw the pictures of a mob
storming the U.S. Capitol on January 6th? I mean, it was a horrible feeling, you know,
never in a million years I thought you would see, you know, in one of the best functioning
democracies in the world, you know, transitions of power are, you know, hallmark of a democracy
working well and, you know, definitely shocked me the same way I think it shocked people around
the world and I'm glad the reaction was what it was and it's a learning moment for all of us and
I think, you know, you realize the foundations of democracies are essentially fragile and it
takes all of us to understand that and reinforce it every way we can. I think we can find some
twist also in how the public sees the whole issue of misinformation. There was a study published
today that says that something like 59% of the American people surveyed in an online poll
think they support breaking up big tech because of the issue of monopolies, of dominance and
of misinformation. What do you make of that? You know, I think there are real issues,
important societal issues to be debated about, you know, where do the boundaries lie with
freedom of speech? What do we need for a well-functioning society? These are important
questions. No single company can solve them and, you know, and I am glad there are conversations
including potential regulatory approaches to give clearer rules of the road and I think large
technology companies have an important role and responsibility to get this right and, you know,
for us as Google it's foundational to our mission. Every single day in search we get, you know,
billions of queries, 15% of it which we have never seen before and we work hard to present the most
accurate information possible and but having said that, I mean, the internet is so distributed.
There are conversations happening across every aspect of the internet. You see important even
just yesterday when you see the, you know, the conversations in the US around the stock market,
movements in stock, where the conversations are happening shows that, you know, internet is,
you know, disseminating information at a scale never seen before.
This phenomenon is bigger than any single company. It is here to stay and I think as a society we
need to develop the next set of frameworks for us to, you know, function through that and I think
that's the debate we are in the middle of. We take our role as a company seriously but and I think
and rightfully that needs to be focused on the larger companies but I would argue that,
you know, it's an internet wide issue and I think the debate needs to be deeper.
I like the idea of working on a new framework. I think that is something that is probably
necessary and will take a while and the whole issue of freedom of speech, it's a really difficult
question we are facing right now because even though you rightly say that single companies
can't decide about freedom of speech and what is on the internet and what is banned from the
internet, in a situation like the 6th of January and the days after, some decisions had to be taken
so a very practical question, do you think Twitter CEO Jack Dorsey took the right decision in banning
the ex-US President Donald Trump from Twitter? You know, I mean, I do think, you know, platforms
operate in a context and, you know, I think, you know, every platform needs to have clear,
transparent rules about what is allowed, what's not allowed and you have to do it in the context
of what's right for your platform. So, you know, I definitely don't want to comment on another
platform. What I would say is I think collectively all of us were dealing with an extraordinary
moment where we saw, you know, the potential for real harm in the physical world and, you know,
potential for incitement to violence and I think, you know, you know, free speech is foundational.
As a company, we have always stood for free speech but there are boundaries which I think
as societies we need to agree on and, you know, incitement to violence is an example of one of
those kinds of boundaries. So, that's the context in which I think we need to look at the moment
and I think, but I do, to go back to my earlier point, I do think it's important for governments
to, you know, debate this and give clear guidance. The answers are going to vary. There is no one
size fits all. I think it's important different companies have different approaches.
I think that gives people a choice but I think we need clearer rules of the road and I think
there are a variety of approaches governments are looking at. For example, I think ensuring
that policies are transparent, decisions are explained, people have a way to appeal those
decisions and overall companies issuing transparency reports, those are all examples of
clearer rules. I think like we have done in privacy with GDPR, you know, I think there are
advances like that which we need to need to undertake. Google just opened a second safety
engineering center in Dublin in Ireland and it's, as to my knowledge, the first one specifically
focused on content responsibility. That sounds like a small step toward a clearer acknowledgement
that the technical platform and the content provided on it can't be separated, is it?
You know, I'm very excited about it. It's our second Google safety engineering center. The
first is in Munich in Germany on privacy. This is focused on content responsibility,
not just brings our experts, technical and policy experts and our trust and safety teams,
but allows us to engage with other stakeholders including academic, government experts, etc.
You know, content responsibility is an ongoing. We will never be done with this work.
You know, we take it very seriously. It's one of our largest areas of investment and I see that,
you know, I gave the search example in YouTube. It's been our biggest area of focus for the past
four years and I look at many areas and I see the progress we have made, but the scale of the
internet, I already mentioned, you know, how the context keeps changing. You know, something like
COVID didn't exist a year ago, right? And so on a given day, the nature of information and the
problems, you know, these are constantly evolving and so I think that's why we are setting up this
center. It gives us a way to engage externally and, you know, I don't think these are societal
level problems and you're going to need multiple stakeholders. You need public-private partnerships
to get this right and I think for us, we want to set up organizations and a way by which we
can engage, particularly in Europe, so I'm really excited for it. The political year is expected to
bring something like a paradigm change in terms of tech governance. What do you anticipate from
the new Biden-Harris government administration? You know, I think, you know, it's definitely
early days. You know, as a company, we've always been committed to engaging with the government
regardless of an administration. As a company, we want to do our part. So, for example, if it's
COVID, we've been focused on what our role can be in economic recovery, supporting small
businesses, helping people gain digital skills. You know, we have announced initiatives around
racial equity, vaccination, how can we help in vaccination. So, we are going to engage on
areas where we can find common ground and advocate for it. Sustainability has been an important
area of focus for us. I've been encouraged, you know, along certain dimensions. I think the
Transatlantic Alliance is something that needs to be, you know, invested in and restored and
I think it's really important. When I look at challenges like sustainability or AI regulation
and safety, we are going to need, you know, more global cooperation and so I'm encouraged
by the early signals from the administration. But the question of how the Biden-Harris
administration will deal with big tech is, of course, not just about cooperation. We are
talking hard facts there in terms of competition law and issues of monopolies. How do you prepare
for a time that might be much more uncomfortable for you as a company as well as other big tech
companies? You know, I've long, you know, held the view that it's appropriate for, you know,
democratic societies to hold large companies accountable and this could, you know, this is
not new to us and the administration change, you know, we have had, you know, either processes and
procedures in the context of previous administrations as well. I'm glad there are, you
know, as a company, we've always been focused for 21 years on making sure we are innovating,
contributing value to society, helping drive products and services for users in a beneficial
way and as part of that, we will engage constructively. There may be feedback on
changes we need to undertake and, you know, we feel it's our responsibility to do that as well.
So, I don't necessarily see that as a new thing. It is something definitely that comes with scale
and I, you know, I think it's our duty to engage with regulatory bodies, with, you know,
democratic societies and answer questions they have and so I see it as a natural part of our
evolution. Okay, we have talked a lot about stakeholder capitalism at this World Economic
Forum. It's not a new concept, but I think especially in these days when we kind of
recalibrate the values of how we work together, how our economy thrives and all these kinds of
things, that's a concept to be reconsidered and I would like to ask you how you combine
that a few days ago Google workers across the globe announced Alpha Global, an international
union alliance to hold Alphabet accountable in a way. Do the employees have the feeling or do
you have the feeling the employees might think they have to take the matter of stakeholder
capitalism into their own hands now? You know, I think, you know, as a company we've always held
a view that, you know, that employees, you know, are an important constituency for us. You know,
we deeply care about our employees. We give, you know, giving voice to employees is a core
part of our culture and we engage with many employee resource groups in many forums,
including in Europe. You know, we've had European worker councils for a long time and so, you know,
I think this is not new for us. I think listening is an important part of our culture and I do think
we want to work with our employees to drive progress together, be it on areas like sustainability
or diversity and inclusion and so I think I continue to believe that, you know, companies
which are able to engage with employees and involve them, you know, you know, it's a strength
of the company, I feel, and so that's how I think about it. I would like to press you on that a
little bit more with one more question because Google has been struggling with a series of
employee issues these days. There was one with the most public awareness being the
depending on the prospective resignation or firing of Timnit Gebru, a well-reputed researcher
on Google's ethical AI team. How seriously does Google take the issue of the critical questioning
required to develop a responsible AI and how independent can research
in a company be to really address these questions?
You know, I mean, issues around, always AI is one of the most profound things we are working on
and, you know, developing it responsibly is one of the most important things for us to get right.
As a company, we are publicly committed to AI principles. We are, you know, one of the leaders
in industry in terms of the research we do around AI ethics and we published over 200 papers in the
past year or so but I think you're raising an important question which is, I don't think,
I don't see it as company, we are committed to be, you know, doing the best we can in terms of
approaching our work in a responsible way and pursuing AI ethics but I think there is an
extraordinarily important role for regulation, for, I've said AI is too important an area not
to be regulated, for independent institutions, academic, other non-profit institutions to do
this and I don't think companies alone can get this right. I think part of the reason you see,
you know, a lot of debate, I mean, we engage as a company, we are a lot more transparent than most
of the companies and so you do see us in the middle of these issues but I take it as a sign
that, you know, we allow for debate to happen around this area and we need to get better as
a company, we are committed to doing so but I, you know, I look at the state of how we approach this
and, you know, I am confident about the effort we are putting into it and our commitment to do
better here over time. Maybe let's use the last minute to take a look at some new technologies
and how they might help our economies, our societies to thrive. At the end of November
last year Google's DeepMind, you mentioned that at the beginning of our conversation,
announced a gigantic leap in solving protein structures. The deep learning software AlphaFold
manages to accurately predict protein structures from their amino acid sequence. Can you give us
an example of the practical value of this breakthrough? What will we do with that?
I mean, I think it's today, you know, drug development is an extraordinarily involved
process, you know, both in terms of time and resources, in terms of people and money and
what something like AlphaFold gives the chances for biotech pharmaceutical companies, anyone in
the drug development process, to become much more efficient and streamlined and to find the right
protein structures for a target problem and so effectively streamline the development of a
solution. And so I view this very profound but, you know, it's one step of a complicated drug
development process. But that's how advances happen, you know, and when we made progress
with messenger RNA technology, today we are deploying it in the context of a pandemic. So
I view it as a foundational advance and, you know, and I think, you know, I see countless examples
across the company and, you know, about 18 months ago we demonstrated quantum supremacy
as a company and I think it's another example of a foundational advance which will help us
maybe model, you know, climate better in the future. And I think progress takes time and,
you know, I think all of the internet today is built on some foundational technologies which
were developed 30-40 years ago and so that's how we achieve progress. I still don't like the term
quantum supremacy. It sounds like a very military phrase. I prefer quantum advantage.
But let's stick to what might come out of that. When will this technology be available for a
broader market quantum computing? I think what you will see is over the next three to five years,
I think, you know, a few places will give outside companies and developers
to, you know, a quantum cloud, if you will, for them to, you know, tackle early research problems.
So we are definitely early in the cycle. I think, you know, the demonstrations, you know,
show the potential of what's possible and that's why we are all excited about it. But I think you
are on a 10-20 year journey of really bringing it to, you know, actual practical use. But the
next three to five years will be exciting because, you know, you can harness the creativity of the
world. That is, you can let other developers, you know, test quantum applications by providing
the quantum infrastructure which we are building with APIs to the outside world and I'm really
excited for that. Are you worried that Chinese researchers announced at the end of last year
to have reached quantum advantage as well? Will that be the next battle of the tech titan nation
somehow? There is no doubt to me, you know, across the world there are going to be progress in areas
like AI and quantum. I mean, these are foundational areas, you know, China with its technology
resources, the investment from the government is definitely going to make progress across these
areas as well, you know, which is why, you know, to my earlier point, we are going to need
better global frameworks for AI safety. Just like to tackle something like climate change,
we have the notion of a Paris Agreement. We need countries to come together because no
single country can solve a planetary issue. I think issues around, you know, as technology
progresses with AI and quantum computing, we're going to need additional frameworks like that to
ensure safety for the planet and so I think it's important to acknowledge
there is going to be global progress and I think, you know, we should think early and
come together to solve the bigger, longer-term safety issues from these technologies.
So we somehow need a global quantum regulation framework. Who is going to provide for that?
It's not just quantum, it's around, you know, AI safety and, you know, we need to make sure that,
you know, just like in the past, we have had other frameworks globally and they are definitely not
easy to do. It needs willing countries and governments and, you know, enough set of
committed people to drive progress here and, you know, I think it will be multi-decade efforts.
You know, I expect, for example, I hope in G20, in OECD, I mean, these are all the forums and,
you know, there are discussions happening around this and I think, again, you're going to need
public-private partnerships to drive all of this forward.
Let me challenge you in one more meta perspective regarding quantum computing.
Quantum computers work by the principle of entanglement from quantum mechanics,
which means that each particle of the pair or group cannot be described independently of the
state of the others. What I thought is, isn't that kind of a wonderful metaphor of the world
we are living in and might teach us something about humbly accepting complexity, accepting
ambiguity in the state of the world we are living in right now?
Ah, you know, definitely, you know, quantum computing is one of those wonderful things in
which the more you spend time trying to understand it, it's humbling and you get very philosophical
in nature. No one other than the great Richard Feynman said anyone who claims to understand
quantum computing doesn't really understand it and we get the exact quote wrong, but
I think there is definitely something about the notion that the physical world as we perceive it
doesn't represent the underlying reality and as a humanity we have a long way to go
before we understand the true nature in a deeper way. What I am excited about quantum physics and
quantum computing is it gives us a chance to get closer to that understanding one day.
One last question. Looking at your background, what is that little dinosaur right behind you
on the shelf? Oh, you know, I think when we were developing Chrome, the browser, you know,
when there was no internet connectivity, you could play the dino game. And so it's something
we did in our early days. And so it's a little bit of an Easter egg for me that I carry with
at times. I thought it's something like never look back to the things that have been
extinguished forever, but look to the future, maybe. I think that's a good perspective to carry
too. Yeah. So we're running out of time. Thank you very much for this conversation and the insights
and ideas. Sundar, it was a pleasure. And thanks to everyone out there in the virtual realm. Thank
you. Thanks, Miriam, and thanks to the World Economic Forum for organizing this. Appreciate it.
