Podcast: "Ok, Google: Give me the World"

TRANSCRIPT: "Ok, Google: Give me the World"

Fei-Fei Li:

AI is no different from other technologies. It'll pose new challenges that humanity has to collectively be aware of and solve. It'll also create opportunities we've never thought of.

Ian Bremmer:

Hi, I'm Ian Bremmer and welcome to the GZERO World Podcast. I'm host of the Weekly Show, GZERO World on Facebook Watch. In this podcast, we share extended versions of the big interviews from that show. This week I sit down with Fei-Fei Li. She's Chief Scientist of AI at Google Cloud and Director of the Stanford Artificial Intelligence Lab. She's a staunch advocate for bringing more women, especially women of color to the table in the tech world. Today, I'll ask her about the future of artificial intelligence, how countries should deal with the issue, and whether machines are really going to take over the world. Let's get to it.

Announcer:

The GZERO World is brought to you by our founding sponsor, First Republic. First Republic, a private bank and wealth management company. Imagine a bank without teller lines where your banker knows your name and its most prized currency is extraordinary client service. Hear directly from First Republic's clients by visiting firstrepublic.com.

Ian Bremmer:

I am deep in the bowels of the googolplex here in the Bay Area with Fei-Fei Li, the Director of the Stanford AI Lab and Professor of Computer Science there, but also now Chief Scientist of AI and machine learning here at Google Cloud. Fei-Fei, great to be with you.

Fei-Fei Li:

Great to be here, Ian.

Ian Bremmer:

How much of the game changer that Elon Musk and Stephen Hawking talk about AI is real right now in your view?

Fei-Fei Li:

AI is a deeply impactful technology. I think before Elon Musk and Steve Hawkins talking about AI, a lot of technologists and thought leaders have recognized the importance of AI as a technology. It's a new way of thinking about computing, and it has the transformative power to change many different areas of our industry. And so I deeply believe the importance and impactfulness of AI.

Ian Bremmer:

We talk about a lot of these technological changes as hockey sticks. They come along for a long time, we know that they have potentiality, and then all of a sudden they explode. You are in the middle of this in a way that almost no one else is. Where are we right now on the hockey stick?

Fei-Fei Li:

AI as a field, Ian, is actually very young. It's only 60 years old. If you think about some deeper and more classic fields of human science and technology like physics, chemistry, they have a way longer history. So first of all, it's not surprising that AI seems to be a new thing, new kid around the block. Second of all, after 60 years of cumulation in terms of basic science research and research and development in the industry, we are genuinely seeing the first wave of AI explosion.

Ian Bremmer:

The media focuses on the big wins of computers against human beings. It's when Kasparov goes down in a chess match, it's when the world Go master suddenly says, "I don't know how they figured it out." But are those the real stepping stones that we should be focused on? Or are there other things that aren't as sexy, but are much more society changing?

Fei-Fei Li:

If you look at the past 20 years, I think the bigger technological impact and change came from the internet. So I think there is a little bit of a mismatch between a single game of machine versus men, versus the real deep transformations technology or AI brings to the world. So personally, I think AI is already making deep transformations. It's impacting the way people are thinking about communications. It's impacting the way people are thinking about transportation, data analytics, healthcare. These are the deeper changes that I feel inspired by.

Ian Bremmer:

So we know one of the things that people have been talking about the most, and sometimes in the most dystopian ways, is about how AI is going to destroy the workforce. Now, rather than me asking you that question, let me ask you something a little different. You're a mother, in addition to being the chief scientist here. If you think about children and what skills we need to give them, how they need to think about tooling themselves to be more effective for AI empowered workplace, what is that? What do we need to be doing? How do we think differently in that environment?

Fei-Fei Li:

So I get this question a lot.

Ian Bremmer:

Well, not from me.

Fei-Fei Li:

Yeah, I guess I'm a mother and an educator and I'm a AI technologist, so a lot of friends ask me how to prepare the next generation. I just think personally for myself as a parent, the foremost important thing is actually bringing up my children to be the best human beings they can be. And this becomes particularly philosophically important in the age of machines. I really believe technology is here to make us live better and be better humans. So there's no independent goal or value of machines, they should serve the human values. And from that point of view, that humans through our civilization, we never cease to innovate better technology to better our lives.

Ian Bremmer:

You think about technology as serving human beings, and there are an awful lot of human beings in this country right now that would argue that technology has not served them very well. Now, maybe that's about values, the values of the people that are employing that technology. We talk about whether it's fake news or media complex, propaganda, what have you. Now, if I think about the companies and the private sector actors that are driving AI in the United States today and some other countries too, fiduciary responsibility, short-term profitability, male dominated culture. I wonder, how does that make you feel about the kind of values that we're likely to see employed as we develop AI?

Fei-Fei Li:

Yeah, good question, Ian. So I do believe our technology reflects our collective value. So some of these issues you bring up will deeply impact our technology. It's not just AI, AI being the currently most salient technology. So as a technologist myself, I want to advocate human-centered technology, human-centered AI, and this is really important. For example, whenever we talk about automation, the word automation almost automatically equals to replacement, and that's just not the case. Automation should be augmentation for human capability, should be interaction, should be collaboration, should be assistive. So this is the kind of technology that is human-centered, is focusing on increasing the productivity, safety, quality of life of humans.

Ian Bremmer:

The headlines around AI seem to be either about I'm going to make a lot of money, or it's about the robots are taking over. I don't necessarily see the direction you're talking about.

Fei-Fei Li:

I definitely wish that rational as well as well-rounded discussion would happen. I'm also lucky to be surrounded by people who care. My colleagues at Stanford and Google are developing technologies that are trying to reflect the kind of values. We have colleagues developing AI for education models, machine learning for sustainability applications, healthcare, assistive robotics. We have colleagues looking at fairness in machine learning. So I am surrounded by those people, but outside of my immediate colleagues at large, I would wish for a more engaged conversation on these topics.

Ian Bremmer:

Do we think that the changes that are coming in AI are going to lead to more structural inequality in the world today?

Fei-Fei Li:

That's a great question, Ian. Technology advances have in history caused pain in human society, and you know better than I do, the different eras from steam engine to electricity to cars, to computers, all these technologies have caused pain, and they also improved human lives. Even 50 years ago, mortality rate for mother and newborn babies are way higher than today. And look at the access to clean water, to electricity it has. So it always come with these two sides. So I think AI is no different from other technology, from that point of view is that it'll pose new challenges that humanity has to collectively be aware of and solve. It'll also create opportunities we've never thought of.

Ian Bremmer:

Some have a view that technology is necessarily neutral, and what you put in determines what you get out, it's a multiplier. And others would say that, well, it depends on what kind of technology, how it's structured, and that will incent certain types of behaviors. We already talked about the values that come in, but even leaving aside the values, just certain types of code, a certain type of filter, a system that says, I want you to spend money or I want you to engage in this narrow way. And I wonder if you think that society is ready for these changes?

Fei-Fei Li:

AI is a technology that is developing rapidly. So how ready are we? It's a question mark we all have. I don't have a very informed answer. As an educator, I try my best to inform as many students as possible to prepare us for this, by both building more better technology as well as understanding this technology. I will ask you, were we ready for nuclear technology?

Ian Bremmer:

I think the answer's pretty clearly no. And that's kind of what I'm getting at, is that when political systems are fragmented, so we're entering a world right now where the United States is not doing a lot of global leadership, but companies, private sector companies not really aligned with the government, including the one right here, are making these extraordinary breakthroughs. Not sure how that aligns. And as someone who comes out of the academy, a scientist making these decisions, I mean, Oppenheimer was clearly very conflicted about the nuclear bomb, and in part it's because the Americans were fighting the Nazis that he thought this was a great thing. But the Americans aren't fighting the Nazis right now, and so I wonder how Fei-Fei thinks about whether or not this is a technology you think should be embraced and pushed as fast as humanly possible, or if actually there are real dangers around doing that?

Fei-Fei Li:

So I definitely don't think any technology should be pushed as fast as possible without thinking and without care. I've been advocating human-centered AI in terms of thinking about the type of AI technology we develop as well as AI's human impact.

Ian Bremmer:

So a couple other things I want to ask you about. One is, of course, the Americans are not the only game in town when it comes to AI. There are other countries working on this, the Brits, the Canadians, but particularly the Chinese. And the Chinese government has decided that this is a moment for them to be all in state-wise on AI, massive state 2030 plan for developing AI. Were you surprised at the level of ambition? Are you impressed by what the Chinese are doing right now in the field compared to what you see in the United States?

Fei-Fei Li:

I'm definitely impressed. So as you know, natively, I grew up in China and got my education in America. So when I look at what my friends and colleagues are doing in China, I'm very impressed by the passion and aspiration in AI. I have many students who have come from China to study AI overseas. In international AI academic conference, we see a lot of great Chinese authors and institutions publishing good work, so I'm very impressed. I'm also very impressed by the government level of support. I'm not that surprised because I think there's a great drive in China for the past several decade now to make technological advances. It's a country that is known to embrace technology. Look at what China did for mobile technology, for social media, for internet technology. It's really quite phenomenal.

Ian Bremmer:

Are there areas right now of AI where you would say definitively the Chinese are in the lead, or are likely to be in the lead going forward?

Fei-Fei Li:

I think China has room to grow in basic science research in AI. I think what China is getting really good at is where there's good user use case scenario. For example, China's social media like WeChat or Alibaba.

Ian Bremmer:

Because it's more all immersive.

Fei-Fei Li:

Exactly.

Ian Bremmer:

They've got more data, more people.

Fei-Fei Li:

Exactly, exactly. So those areas, I think China is making phenomenal progress.

Ian Bremmer:

So you made one comment that I thought was really interesting when you said we really need to get away from the hoodie culture, even though you have kind of a reverse, very dressy hoodie going on there today. What did you mean by that?

Fei-Fei Li:

So this has come out of an interview that Melinda Gates and I did a few month ago. We both have come together last year to really start to worry about the lack of diversity as we're going through yet another great technological transformation and renaissance. And what I mean is that we thoroughly lack the diversity that represents humanity in the development of technology. We lack that in the first phase of computing, and we're lacking that in the development of AI. And AI technology is so critical for all of us. We talked a lot about values, we talked a lot about AI empowered products and services that would serve people, and we talk about policy, we talk about impact from geopolitical to labor market, all this. We need the balanced voice, we need all of the voices to be at the table.

Ian Bremmer:

Final question for you, and it's kind of a big one, but it's the one that you hear so much about, which is the people that are talking the most dire expressions about the future of AI warn that this is the end of humanity. And I have to ask, when you really are working cutting edge in this field, how much do you worry that you are actually opening Pandora's box?

Fei-Fei Li:

So Ian, I have a 22-month-old child, and I look at her development, like her visual ability, her manipulation, her navigation, her speech, her emotional communication with me. And I look at today's AI algorithms, I think we are very, very, very far from creating the machine that can come that close to that capability. So from a technology point of view, I think we have a long way to go to make machines that are in the realm of Hollywood's imagination. But having said that, I think the way I think about Pandora box is that whether it's AI or any technology, it could become a Pandora box. Humans have this capability of making mistakes with technology or anything else, and we should be on the guard. We should think carefully and we should be thoughtful. We should engage government policy makers, thought leaders. We should be doing all that.

Ian Bremmer:

Fei-Fei Li, not about to be replaced by a robot.

Fei-Fei Li:

Thank you, Ian. You too. You won't be replaced by a robot.

Announcer:

The GZERO World is brought to you by our founding sponsor, First Republic. First Republic, a private bank and wealth management company. Imagine a bank without teller lines where your banker knows your name, and its most prize currency is extraordinary client service. Hear directly from First Republic's clients by visiting firstrepublic.com.