New Work Summit 2018
February 12, 2018, New York, USA
New Work Summit 2018
Video
(Is the Future) Made in China?
Available
In cart
Free
Free
Free
Free
Free
Free
Add to favorites
24.3 K
I like 0
I dislike 0
Available
In cart
Free
Free
Free
Free
Free
Free
  • Description
  • Transcript
  • Discussion

About speakers

Kai-Fu Lee
Chairman and CEO at Sinovation Ventures
Rebecca Blumenstein
Deputy Managing Editor at The New York Times

About the talk

China has laid out ambitious goals to advance AI innovations such as robotics, autonomous vehicles and medical equipment. One of China’s foremost venture capitalists and tech trailblazers shares his insights on the global competitive landscape, as well as the human impact of this coming surge in AI applications.

Share

KY to thank you so much for joining us who made the trip all the way from BJs and we're just so honored to have you you're one of the world's foremost expert on a i and also Tech in China former head of Google Microsoft in China and and you've been you've been writing a lot about Ai and both its promise and and some of its perils but it wanted to start out not many people know or maybe a few people know that you also invented speech recognition technology when you were at Carnegie Mellon to say that was that was that was early stage AI

time we saw there was the first working speaker independent continued speech recognition and we're very proud of its thought that would change it to the future of computer interface in the next 10 years or so. And in fact John Markoff from New York Times flu all the way over here to write about it. It's 1988. so what took so long because yes, I think we got overexcited that the lab demonstration work. It didn't deal with noise problems the way Amazon Echo did it had a 1000-word

vocabulary not the internet vocabulary and the accuracy wasn't as good because there wasn't as some there wasn't as much data to train it the Albertson's in fact were very much to lift machine learning-based and very comparable to keep running. I'll maybe you think of it as primitive version of 1/16 reality. What makes you so confident that the age of AI is firmly here today using Amazon Echo algorithms are doing better than Insurance in terms of default rates. They're doing better for them than previous algorithms

for Credit card fraud but I think there's a lot of empirical proof that they work a lot better than previous algorithm that pipe and people and the primary cause of that is a huge amount of data, which is the fuel of AI that is being gathered Amazing Race is almost like the internet was created and then we communicate with the information is uploaded. We communicate and buy things and talk to each other and all the state of growing rapidly has made the AI possible which Arthur mentioned talks about

China and the competitive advantage that I wouldn't say dominate. I think I probably China and us will each have 50% chance to lead the world. Common a certain areas. So if we split into multiple types, I think for internet I think China will catch up rapidly and Lead because of the amount of data for business AI us will leave for a long time because of the much more advanced state of digitization and structured data and data warehousing for perception date a I I think China will

lead in terms of using a lot of cameras microphones to capture information. But America will probably lead in the advanced interfaces and then. Onimus vehicles and Robotics. It's really am not who will lead because u.s. Is way ahead in technology, but China has no policy advantageous. So coming back to your question about what are the time that the advantages and disadvantages all clearly us leads the world in AI technology if you take the top hundred People at least 80 are Americans China has maybe two so that that's something that's a very clear but at the same

time China has several major reasons propelling it forward for AI one is a huge amount of data. The number of users is dramatically larger three times more for mobile users. Mobile payment is a prevalent in China. So all that data and all the applications that enables 700 million people it is the number of transactions is on the order of fifty times larger than America you look at other day that has been captured shared bicycles 300 times food deliveries 10 times. So we're talking about the order of magnitude more data plus it's an environment where more people are

willing to trade some degree of privacy for convenience. The third reason is the powerful VC entrepreneur ecosystem. Obviously, very strong here and there is trying Chyna butt in China. I think they're Is a much stronger hunger hard work desire to success all in kind of attitude. And I think that will lead to faster the speed and the last reason is sap course government support China came out with the plans. They Council in last year that describes the China deserves to be the Hub of AI Innovation by 23rd.

And a lot of resources are being put behind City level project in provincial-level projects to subsidize. Aiai. Research are people AI Investments. And also I think the Chinese government is a what I would call it techno utilitarian government. Let the technology launch see how it works. Is that a problem fix it and an AI needs exactly that that launched it. I mean the success of AI Acme. Launching as fast getting today that using the data to be fuel to make the technology work even better and better that's very much helpful today. I now of course, I'm I'm not

saying it's not important to consider. The other issues reliability issues ethical issues FiOS issues. But if you have a techno utilitarian approach you are launching its first and then trying to understand and fix the issues, I'll ask you go long. Do you think that the US has underestimated China and part because when they think about tech in China a lot of us think about censorship and you know, some of the more Western views of the internet Oh, I think this phone probably does not and your readers probably do not because I right enough outfits for you and do nothing of you

but I think buying large. Yes. I think there is a perception China's a copycat. There's a perception that Chinese companies can even lead in China at all do different protectionism and there's a perception that a ISO Advanced China has to be way behind and I'm all Street perception is are wrong about some of your impressions of the 10th, which kind of fundamental in terms of what the differences are between Silicon Valley in China today. Yeah, we were able to

see a lot of things. I think the entrepreneurs are very excited by You Know You Remember lyrics vision of the world. They say they saw the singularity University the X prize. I think this song Silicon Valley as this dreamy Visionary place that there's to think big that was very motivating for them. But what was the most common surprise was that we were unable to schedule meetings after 6 or on weekends. They were used to working 997 that is 9 a.m. To 9 p.m. 7 days a week. Some of the laser companies are 996

until I take if Sunday's I'm sorry, they're not available. This is what are we going to do? So we've so we took them to see Golden Gate Bridge and all that, but I But I think they they they would prefer to keep working. I think that's hunger and dedication is what made China strong economically today and and it's some it's basically a fundamental value for these entrepreneurs to work extremely hard and compete very fiercely and design for success by I'm work. It's your view that there is going to be in the black and white of it which we discussed last

night that that that you're more of the side of fundamental change too many jobs as we know them because we are Venture capitalists and we talked to a lot of companies some certainly work on symbiotic human machine making each other better, which is what a lot large companies a lot of large companies like to emphasize and we acknowledge. That is true. I think a doctor armed with a very nice. Adjudge armed with a very nice tool will make better judgments and provide Better Health better Justice Etc. Same with lawyers and teachers there

many professions in which the symbiotic human AI is true. I also acknowledged it will be always be new jobs created but I think the issue we face is as a venture capitalist were talking to companies that are building algorithm software algorithms that do a better job than customer service Representatives that do a better job than telemarketing that do a better job then loan approvals that do a better job in portfolio management. So if you extrapolate the compound that with the pace at which AI is improving and also consider that's not all of them are

141 job replacement. Some are basically completely changing a paradigm shift in the field. The number of Replacements will be substantial. And those Replacements cause several major issues, I think historically we've always dealt with technology replacing jobs by believing the symbiotic nature be leaving our creativity in creating new jobs and you're working few hours there lots of solutions to that. It shouldn't be a problem. But this time it's faster and it's pure software. It's actually not manufacturing jobs that will go first. A lot of people

think it's manufacturing jobs. So China has a bigger challenge. I don't think that's the case because manufacturing jobs requires robotic advancements which are much slower than software Michigan recruiting jobs. And then I think the spies are small disagreements with various people who do research we all agree that a job transition is needed and that transition may be a tough one because if you You can't just say okay, you're not going to do telemarketing to customer service because that job will be gone too and then you move

on and then so on what are the sustainable jobs that will be her to last. We think those probably have to do with creativity complexity dexterity and compassion and not everyone's going to be suitable for that transition, which will be a shorter. Of time than previous technology advancements. So that's something I have been speaking about because I think we as a society need to be aware and prepare for this type of transition. How soon is this unfold? I think the AI algorithms will be able to technically do

the tasks of a lot of human job stuff in the next 10 to 15 years, maybe 50% but it doesn't meaning of 50% unemployment rate is just means tasks of jobs can be done cost-effectively technically effectively by the machine. It's also doesn't mean that the companies will necessarily layoffs the people they can choose to stay with the people and not automate but I'm just saying technically feasibility is 10 to 15 years away and that at least points at a ceiling of a Max impact now many studies. You hear from James from Mackenzie who will think that

translates to you know, how much smaller number of job losses but nevertheless it's not just job losses. It's it's a loss of meaning, you know, people most people in the society believed in hard work and you'll be rewarded. But they are now saying hard work. I will not be rewarded if it's routine and replaceable by I so, you know, there can be social welfare that can be retraining but I think people fundamentally are going to feel a loss of meaning and then with the loss of meaning that can lead to depression and other instability is it also

requires a retraining but it's not easy to take someone from a let's say a assembly line whose job is replaced and say okay you're going to be a caretaker. It isn't clear that position will be easy. Those are the compounded issues a bunch about this in the US. I don't think there's too many easy solutions is the Chinese government worried about this. It doesn't seem to be a good recipe for China and either way. It's very comfortable. I think the amount of impact was very comfortable to us and China. In fact, I think we should work rather

than discussing who stronger in the Iowa should all come up with ideas from The whole world. I think there are many examples. We can all learn from you know. The craftsmanship of the Swiss and the Japanese and then the philanthropy and volunteerism in other countries and so on but to answer your question, I think the Chinese government is not much more focused on this compared to the American government. I think both should be more focused on the other hand. I think the Chinese government has had a stronger records dealing with job transition before

a state-owned Enterprises shifted to private Enterprises as automation created the made-in-china manufacturing those were very much with the government job using policies and two diverts are types of types of work to other types of work. So I would still say this by both governments not being sufficiently prepared. The Chinese approach is naturally a little bit easier to Apple cocktails to make this a less heavy bloom. I could be helpful in eliminating us but there's no research that shows that actually face facial recognition is very successful with

white male faces, but not people with darker skin. Could we could we be looking at 8? Well, the humans are clearly the most biased there's plenty of evidence. There's a study on Israeli judges who are more harsh before lunch because they're hungry. And so those are irrational reasons that machines would never be subject to the machines should be trained on the drive data on the right outcomes if your train if you train a day I based on human markings than the human bias is more likely to stay. So if I have a human loan officer Mark good good Prospect for loan bad Prospect

then. Bias what will continue but if we train the system based on the actual outcome, right he or she returned the money or didn't default to her didn't then is more objective training. So I think that by itself will remove some of the bathroom and behavior, but you could also stay well what if a certain aspect, you know gender Grace age predicted higher luck with the phone rang while you aren't you can always found balance the day then right you can make sure it has enough samples of each type of data. If you really felt you this was so I'll let's say if gender discrimination, you could

remove the male-female data from the training data. So there's so many things we can do with making machine learning less than less biased but that would humans you can't require someone spring. So I think things will only get better for AI same is not true with humans are predominantly the AI Engineers. There's a huge gender and racial imbalance and Oh, yeah, I think the balance in AI is worse than computer science and computer science is worse than many other domains. I think we have to to to

work hard to make sure the pool increases. I personally Mentor Mentor several undergraduates at Harvard and other places where women these are the little things that we we can do and I think we should not probably start early with understanding who have the interest and aptitude to go into Ai and explain the impact and fun in this environment. So that will have more more interest. I think only increasing the pool and making sure people don't make assumptions about this this

and not get into it and there are many very good Role Models fei-fei leaves and outstanding one of the best Before we open it up for questions. I just have to ask we're sitting here in Silicon Valley. What do you think about the prospects of Western companies in China? You run Google and Microsoft in China. Google is recently announced a bit of a four Ray back in his Facebook ever going to be able to get into the mainland generally talk about the wrong issues about I'm happy to cover the subject

people generally want to talk about the protectionism allowed or censorship in those issues. And yes, I acknowledge. Those are I have been issues but I think the major issue today for these companies are the following are number one of their products are not competitive messenger is a much worse product in WeChat. How can they write Amazon with grape in America? But Amazon China is substantially worse than how about JD and tmall and how can can they say succeed? I think. Product the payment, you know Apple pay is much narrower and much harder to

use then then we Cherry Valley peso. The American products are simply uncompetitive in the market mean in competitive. Bothan not localized right for the local nice and also just on the product product feature for feature level not competitive. That is the fundamental issue. The second issue is the multinational companies are cannot hire people to lead and run the China operations anymore. The typical head of a multinational company in China surprisingly is someone who ran a European country that moved to run China because that's a larger country and

another language can communicate locally multinationals don't like hiring Chinese people to run the operations and that's terrible and these people are in scented for one thing and one thing only which is can he or she follow the corporate orders in law? Get a promotion to become a VP senior VP back in the headquarters. They're not concerned about winning in the local market. You put them against these 997 hungry entrepreneurs are going to get eaten for lunch. And the last problem which is very big is the American companies are no longer at a hiring advantage in China. Most of the

graduates would prefer to work for Alibaba tencent one of our portfolio companies or Star the company the multinationals on the average are used to be the most desirable place to work now on the average. They're not obviously some people still admire want to work at a company like Google and Microsoft but but on the average that's dropped a lot. So these issues were fundamentally make it very difficult. If not impossible for American company to go one. Last reason is American companies usually theme and it Global platform that they won't change for Chinese needs and the Chinese

companies. Think about when Chinese companies go to Southeast Asia. India South America, they will work with a local partner and adapt a product product make an investment and that is probably the better future approach to take so it lots of issues. I'd like to open it up for questions. I see a hand right there. Jonathan daryanani from cognotion. We do automated with AI based vocational training primarily in healthcare and I was able to go to China to look at some nursing homes this year.

I was impressed with the idealistic side of a I been a number of young Chinese people who thought, you know, you don't have to have democracy have transparency to decrease corruption and increase Merit and they looked at a is the place where Chinese society would realize some of its principles by essentially having an increase of meritocracy and transparency better decision-making was going to improve Chinese Society without political changes. Is that a completely unrealistic and I use you. Do you think that there's anything to that is that something that the young people that you

talk to share? Well, I can argue on both sides. Question. There are a lot of compelling reasons to argue. They're both sides. Here's what I observed today in China. I think it's a very pragmatic Society. So people will use things and experiment with things and there are places where AI is improving transparency and fairness, but may be adopted in the way that wouldn't be imaginable in the u.s. What is AppleOne company has developed the sentencing assistance for judges and it's quite accurate. Obviously, the judge makes the final call, but

I'm with this tool Justice will improve so a can that be accepted here as an improvement or not. I don't know but if you talk about call me back to a bunch of other young people and how they view I think most young people are still following them shoppings Mantra let some people get rich first and and and only a small percentage of China is very wealthy, I think. Do people still follow material wealth as their primary aspiration and that causes them to kind of focus on that and Overlook many of the other issues one could speculate whether that will

change as the middle class emerges but so far, I think people I don't think the happiness or is any lower in China today people feel excited. The country is growing and they're excited. They have the opportunity at least economically. I think that's the general feeling. Thank you Kim Goldberg from UC Berkeley. I'm curious about we've been hearing a little bit about this new concept of one belt one road and it's a mysterious name too many of us and I'm curious if you can shed some light on it. I'm really not the expert on the issue. It's about

reconnecting China to to Southeast Asia to Middle East. And basically, I think it's related to increasing economic collaboration with countries in those areas. And I think it's also related to increasing the alliance and Affinity with those countries. I'd like you to join me in thanking kaifu for a very provocative discussion.

Cackle comments for the website

Buy this talk

Access to the talk “(Is the Future) Made in China?”
Available
In cart
Free
Free
Free
Free
Free
Free

Access to all the recordings of the event

Get access to all videos “New Work Summit 2018”
Available
In cart
Free
Free
Free
Free
Free
Free
Ticket

Interested in topic “AI and Machine learning”?

You might be interested in videos from this event

March 11, 2020
Sunnyvale
30
197.46 K
dev, google, js, machine learning, ml, scaling, software , tensorflow, web

Similar talks

Jim Breyer
Founder/CEO at Breyer Capital
+ 1 speaker
Deep Nishar
Senior Managing Partner at SoftBank Vision Fund
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free
Andrew Ng
Founder and C.E.O. at Landing.ai
Available
In cart
Free
Free
Free
Free
Free
Free
Greg Brockman
C.T.O. and Co-Founder at OpenAI
+ 1 speaker
Terah Lyons
Executive Director at The Partnership on AI
+ 1 speaker
Available
In cart
Free
Free
Free
Free
Free
Free

Buy this video

Video

Access to the talk “(Is the Future) Made in China?”
Available
In cart
Free
Free
Free
Free
Free
Free

Conference Cast

With ConferenceCast.tv, you get access to our library of the world's best conference talks.

Conference Cast
551 conferences
21655 speakers
8015 hours of content