Are we already enslaved to AI? With Nicholas Young
33:56 Share

Hey, daily china is produced together with our friends at radio, this awesome independent media platform. If you're interested in culture and innovation in china, you should definitely check out radio china dot com. They'll give you inside look into everything from china's underground music scene to bike sharing. That's R A D I I china dot com.

I mean, china recognized that they need to evolve for being a low value at the economy, moving into advanced manufacturing. And the only way to do this is use technologies like robotics and A I to enable their economy and their companies.

It's a super White term. Can we just try to narrow down first? I think .

that's a really good question because the narrative that forms around A I is on this leading or just one part of ai. So that doesn't sound all that sexy, but it's essentially a system that interprets and learns from data and then applies those learnings to complete test.

Come considering eva is a wave on holiday, I have a special guest with me this week. Nick is one of the great examples of all the foreign super smart talent that now have moved to china in order to pursue opportunities here. Previously, he was working with A I at the e commerce company. J, D, welcome to dig china. A D, cast fast ating .

second .

together .

with radio. I'm either.

i'm jacket and i'm tom.

Appoint to various studies, china in the street is now affect the lowest in the world. You may know their messaging, I told the chess chinese outbound to M A chinese properties are by internal companies of being the case. Probably never china are.

Show me, yes, it's state. It's clam. The apples major deal over in china, your chinese tech giants meeting at eight point six billion dollar access by a major stake and super sale, forty two point three billion dollars in sales locked by a chinese commerce site in .

one wild day.

Where do I start? I actually grew up in america. My hometown is new york. And so I actually coming to china, and to me wasn't like a lot of other chinese americans coming back and reconnect with the roots.

For me, I was vastly new place because I actually, for generation chinese, and I thought, you know, we going to spend a summer here, are teaching english, you know, improving my chinese a little bit. Subsequently, uh, I ended up changing the entire direction of my life because I was so inspired by what I saw out here. I most recently was working at J D dot com leading their A I innovation team.

And before that, I was actually out in north east china running my own company that brought people from around the world, north korea, and the education programs that we ran in north korea. We used to raise money to support scholarships to bring our queen students out of the country. So that was something that I did for forty five years.

Very different from the A I C um spend time working uh you know a top design consult the C R here went back for grad school um at stanford and I didn't be there. And um you know I never really thought that I would stay so long, but now i've never really given any second thought to leaving. So it's something that has slowly come to from a central part of my identity. I can go born to that as well. But it's just a very, very exciting everything I see here, which really makes me want to stay.

And now you're basing shenton working with the printer company coracle.

correct. So my code venter, and that we started caracal the middle of two thousand eight and the idea that we saw was that, uh, connecting companies from europe, especially in a lot of the start of hubs in europe, which have incredible teams, incredible products, but lacked the the scale of funding that exists in the U. S.

And chinese ecosystem. So we saw this value of these startups to investors in asia in the first instance, in an ultimately around the world. Um and we do this in in a very straighforward way in the first instance.

And we have technology that is able to aggregate and extract information from private companies and transform the information into what we call a story book, which is essentially a immersive interface for, you know, a company's story has its key statistics, it's key product, it's key people. And then these are the the interfaces that we forward to our entire network of investors, especially strategic investors in asia. And then we help the facility take connections between the two.

Today we're actually here to talk about A, I. I did some research on A, I. And my only angle was what are ready? The X, B to c use cases where A, I is supplied. And the only thing I found was within stuff like text, voice, image recognition or video recognition. And then recommendation engine cy way e commerce companies do IT.

Well, I mean, the more relevant question isn't so much what area AI is applied, but exactly how it's applied within that area. You know you if you look at voice, you can see many different use cases for voice.

So you know a classic examples, amazon alexa, when you take that type of technology and then applied to the existing ecosystem that amazon has built around shopping and especially shopping for your everyday things, coupled with you know basically on demand delivery, which is something that J D D C also guarantee. Um it's a very powerful, you know these mart speakers are very powerful technology. So when we're looking at the type of business cases that really makes sense for ai, IT doesn't actually make sense to be begin with. The technology is kind of like going back to basics, figuring out what what the users of the customers need and then think if A I helps to deliver value Better and I say Better, I mean faster or lower costs with the model .

live experience. yeah. So one of the like main things with A I is what you have mentioned a bit before, a machine that by looking at data times can learn itself and become Better and you may be hopefully future smart than human beings.

What I mean to me, this ideas marter also becomes a blurry. If we don't talk about IT in in a more matter of fact way, what IT really means is that pan recognition is something that you do if you have a lot of practice. And so, you know, I believe IT was somehow said that if you want to become an expert, you have to practice something for ten thousand hours.

And so, you know, a human could also accomplish a lot of tasks that a machine could, if you practice over and over over again. And so you take the example of, you know, great chess master will play you know, many, many games every single day and has studied all the great games. 嗯， the A I, you can feed the in a public record of chess games, you know, by the thousands, by the millions, in a very, very short period of time.

The A I can play against itself. You can run simulations over and over again, and continue to learn from those simulations and and make itself even Better. Basically add more fuel to its pattern recognition engine. And because of that, is ultimately outperformance humans. So when there is a task that is repeatedly and when there are rules that are very concrete, very well to find A I excel.

let's take another example. H, let's take the recommendation engine we talked about for the U K. Of, let's say, amazon being able to tell you what you should buy or and guess what you want to buy right now. How does that actually work?

You what we're looking at is how do you process data? How do you process data to improve um the way you deliver services. So the data the company amazon collects every day and you know begins to repeat itself.

The certain types of people will buy certain things over and over again that humans can identify these things, but there are certain things that humans will miss. And so once you know even an A I system that can analyze these patterns at granular detail, you can pick up certain things that, you know, the human, I might not have caught. So for example, you can infer that, for whatever reason, IT maybe a certain cosmetic product, which Normally would be fold to women.

You would imagine that's the conditional is dom the human often fall victim to um is actually very well received by men. And because of that, you know that would pop up automatically in recommendation agent, something that the smartest people of amazon might not thought of because it's forward thinking. You can look at this data and and make recommendations based on the trends in the data that humans .

are much less equip do. So one thing I just can't understand one hundred percent here when he comes to, for example, recommendation engines is does that only actually make the big players larger in the sense that you need that data set in order to deliver an awesome product experience? So does that mean that the start up because of A I never going going to compete with let's a 阿里巴巴 or J， D or amazon because they will always give a more personalize experience thanks to their data sets. why?

It's a good question. And in startups can use A B testing to improve their experience. But generally, when you talk about something like recommendation engines and you're not aggregating a lot of products or a lot of content on just not as many choices.

So is the difference between if you had you know you went into a restaurant and you had a menu that had a thousand different items, if you presented that to a customer, um they would want you to help sort those items and make some suggestions, right? But if you go into a restaurant and there's only ten minutes on the menu or even more humites, it's oma aca, the chef to science or what you get to eat, then you recommendation engine is less important. So let's take the example of a tech brand, which a lot of starts created have very loyal following around a specific brand.

The idea of recommending a specific item to a person is less important than building an emotional connection between customer and brand. And it's building those emotional connections that still there's no, no pattern for that a human creativity is really, really important. I mean, you could apply A I to see.

Well, let's look at all of these other small beauty brands and let's see if there's a pattern that can be identified between the brands that succeed and the brands that don't ultimately succeed. And maybe you can draw some inferences the by a large creative intuition wins in this case because, uh, a lot of these brands are selling products or creating experiences that never existed before. So I think for the smaller companies, you know, these recommendation engines, that mass amounts of data based on their inventory, their G, M V, is not as relevant for a company.

Another example that's very popular globally is a in china called doing and abroad, tiktok former. Musically, they been talking a lot about their allegory, atms in order to figure out who to show certain videos to because your attention spending so short as the users you'll give up after like three, four videos and therefore raw is so important for them to kind of match you to the right content.

So that's very powerful as well. I'm glad you mention that because when we talk about analyzing a product, when we talk to about that with e commerce is actually quite straightforward because the product is catalog is is a light and the light is produced by a company and then there has certain spects. So when you talking about video content, context is much harder to infer, far harder than text.

And so if you have the ability to analyze a video and make some kind of guess as to what that video is showing, and then to be able to classify that video as know, a humorous video, a shocking video, a cute video, and applied these tags to that video, then you can figure out what type of user gravitates towards this type of video, and then show that use of this video. Because ultimately, these social media companies are successful, or I could say they live and die by their ability to retain users. You are looking at the phone for every single waking moment of your life that you're not spending on eating and bodily functions. And actually me that's a scary proposition um because I don't think there is enough for thought going into the type of content that is shown and long term ramifications of having especially Young people, consumers content at the volume that their economic consuming IT there's a lack of um is a lack of morality. This is my own pet ef and I think is a like of morality in terms of what .

these countries are showing kids yeah because at some point the human element or the kind of mission of a culture come in, right? You know, just because there's a pattern that every Young kid must play video games doesn't mean that maybe that he should well.

I mean, this is a very important point because this is, you know, you can have and I begin to analyze patterns, but for what go, that's something that humans can set. And so for a lot of these companies, if the goal is to retain users, then as as your technology is able to analyze more and more patterns of what retains users, you will optimize that. And so you will create content that you know the end state is, is the perfect machine for retaining your users, but that might not be even close to the ideal type of content that is suitable for creating harmonious society or for creating an intelligent society or a society that nurtures values yeah and that actually .

leads a little bit into a topic that I know a lot of people are curious about. So recently, I just saw a top salary list of take companies in china. And basically the five highest paying jobs were A R related.

And you've actually done that type of job. Like what does IT actually mean to work with A I in the company? IT sounds like you're just having a machine. Do all the work for you, you know?

Oh yeah, I mean, there's different levels of A I, of course, but there is definitely work for talent. You know you need the scientists who are are writing the algorithms creating more cutting edge A I. You look cutting edge things with voice in or facial recognition, which is closer to market adoption or it's it's already there, be honest.

And then on top of that, it's you need the tech to to implement the A I within a product. You need a product team to build this technology within something that adapt user needs. And then you know from there is you know you go into the business ops of IT.

But in terms of the A I division, I would say that IT really breaks down into the scientist, the engineers, the product people. And I would also argue design because ultimately, uh, is that he would be to see products. The things that you are creating will only be able to generate insights if they're able to collect data. And only be able to collect data, they're able to maintain usage. And ultimately, designers play a large role into making products to live, to use and keeping people engaged. So I mean, what does the day to day look like is actually quite similar to I mean, if you're looking on the business side, quite similar to take engineering company, except the algorithms that you're writing are are slightly different, then you know if you're coding up a website or you know developing an APP, but it's generally the same kind of thing .

IT feels like to contradicting things because at one hand, you are saying that, you know, we have machines that's going to learn itself and become much Better at voice recognition or check recognition, whatever. But on the other hand, we have a big kind of engineering team behind writing algorithms that steers the ai. So who controls what is a .

good question. The way that I would explain IT to people, because I didn't work on these algorithms at the very technical level, is that you set certain promoters for what you want to optimize. This is what I mention with the social media company. He is trying to optimize user attention.

You said certain parameters for the types of things that you want to analyze, and D A, I technology that you created goes and runs either multiple scenarios or analysis patterns to try to figure out the best way to achieve the goals at the know. The person that wrote the algorithm lays out for itself. And so if it's something around a game that's actually the simplistic example.

The rules of the game are very simple, so the conditions for winning are very simple. The premiers are clear, and so as long as you stay within the boundaries of the rules, the technology can analyze all of the different scenarios for every single decision point and figure out what the optimal way to achieve Victory is. IT become a little bit more difficult when you are talking about things like, you know, how to generate sales and offline retail location.

Somebody walks in the door and let's say you you have some kind of obliged technology that communicate to the sales person how best to talk to this customer. And then the premiers are clear. You want to make a sale, but the rules for what you can do are much less clear. And so in those cases, um I can't say that an engineer would be able to predict how things turn out and you really need to see mom OK. This is a gram effect, the actual act will behavior of humans.

And this actually brings us into this whole discussion about A I right now. I think the elon musk quote, he basically stated that IT is one of the most dangerous things that's gona happen to humankind. IT need to be controlled, legislated, otherwise, you know, we're gonna leave away the power to machines, and they will dominate the world sea. Etra, on the other hand, you know, uh, there are a lot people saying that actual A I technology is just a glorified way to use human labor. So which way is IT?

I don't know if I as president as ella mask. So I do respect this opinion a lot because th Epace o f t echnology i s a ccelerating. So you know the the advances we've made this year will be eclipsed next year and so on and so forth. So at this rate, I don't know th Epace o f w atch w hat I w ill b ecome, but I would say that one of the issues we face around controlling A I is really sending guidelines and premieres for what we want this technology to do.

I get I don't want to keep harding on IT letting social media is the really good example of when you take a recommendation engine that knows so much about you um such that I can manufacture prediction we think about is A I already controlling us because this is something that was designed by humans. We're not even talking about omission A I at this point, but we are already, in some sense, slave to a lot of the the content that is delivered to us through A I recommendation engine. And I know myself is you know these products are designed around exactly these types of insight.

Every person's feed is different. Every person's feed is designed to optimize the amount of time they spend on the platform. And so there's certain examples where we already are quite attached to A I you look at the way we use our phone to do a whole number of things.

I I mention a very simply use case earlier interview around how optical character recognition used to do online to positive checks. Or you know, an online scan of your critical. And you can take a picture, I have a document converted to A P, D F.

These things are very, very simple or converted to like a word document is a very simple technology. And yet we are already our reliance on IT. And so the question is, is there for thinking insight that sets guidelines around the type of technology we can create with A I around certain guidelines that are tied to our values as society, as number one. And I think number two is, do we have the human capacity to think independently about when to put the brakes on? No, I think people who are less optimistic, like you know, must could say that we've already lost the will to put the brakes on these things because we're not even aware of the degree to which we rely on AI powerful technology and and products and experiences.

Okay, but let's drill down on that. What are the actual risks? Let's say we let A I develop freely and all that. What is that worst case?

So I mean, this is for me, um IT kind of borders on the science fiction, but I think is the worst case for me, the scene that I fear that humans ability to think independently. Because why should I think independently when I can optimize my decision making? I mean, think about one's the lesson.

You went driving to a place you don't recognize without using your G, P, S. You mean you might know how to get there, but you know, do you know what the most optimized route is in terms of avoiding traffic? You know um avoiding accidents and red light in these things are already come in place.

We can agree all of this data from cars and you know you use sources all around the world that we just aggregate all that information. You can optimize our travel route. We've ready become completely relying on that.

And so when you look at the way we look at um you know optimizing our own health through recommendations from technology, these are these things have brought efficiency to our life. We don't want to be thinking about these things, but when we surrender the capacity critically on our own IT far more dangerous. And as I said, your examples of this bound social media isn't easy target, but there are other targets as well.

So right now, i'm alizon bunch of people just like spinning bikes to generate electricity or walk, not clean games. The black mare episode. So that's your future prediction. That's where we going to end up.

What I mean is that is part of A I, but it's also living in a virtual world in some sense, if you can game ify the way I interact with reality by creating a virtual reality, and I know longer need to interact with the physical world around me, lets me honest, that already happens. I mean, black mir has placed IT within a setting that looks far more futuristic. But look at the way we live already.

We basically are standing on our phones most of the time. And if we not staying in our phones that we're during its screens, yeah, every surface can be digitalized and and can be transformed into you a digital interface. So already we are completely glued with the screen and we are usually doing things um in response to stimuli or incentives that these companies create for us.

We score because we got apple notifications. We scroll through your news feed or commerce. You know it's a power of response.

You get a little thing and you look at your phone, phone vibrates, you take your even realize this anymore and to think like less than, well, when do I get my iphone in my first iphone, I was late to the game. I got one two thousand less than ten years ago. I didn't even know a smart phone was now I can hardly do anything without IT.

Yeah, that's actually true and it's a really good point. But let me just take the millennial side of this, which is, nick, you're just too old. You understand new modern stuff.

A lot of people do joke and say, old man, because one of the goals that I said for two, that in nineteen is that I should read books more. But look, I I may be old. But here's a thing, if you are looking at a lot of A I technology and I look what it's done for people and one of the major benefits to be at technologies that, that helps us save a lot of time.

Once again, I I go back to another example just for the sake of clarity. And i'm talking about, you know, depositing a check just by taking a photo at that saves time. I don't want to walk to the bank. I love that know if I can walk through A A door just and I want to go stop at the security desk in the, you can, my face, I walk the same time.

But the question that is, what am I doing with this extra time? And I doing something to enrich myself as anywhere time, my kids, my friends, when we can get, you know, pretty pathetic around how we should be spending our free time, is something I should be telling people how to do IT. But I, you go right back to our phone with free time.

I know I do. I have some quiet time to, you know, relax, some question. This is time that I could be reading a book, or talking to a friend, or calling somebody who I haven't talked to, talking with parents.

So I don't call home enough, but that will take my phone. Now, just watch youtube, which knows exactly what I like now, just scary. I didn't realize I was such a charlier son exactly what I like everything about me it's not it's it's not a great reflection in the but um yes, so what do you do with that extra time? That's what I I ask a lot of the people you in my milenio generation, in our milenio generation .

yeah I definitely get your point. And in so many ways, I agree with you.

On the other hand, like I think that's where the issues are right now, right where where it's really hard to draw kind of that line between what gives us efficiency and there for good for society versus, you know, what we do with our time or one is becoming a little bit too efficient than making us maybe acy when we don't want to think, can we just want to wake up and a computer tells us, oh, now you need to do the following three things in order to outlook next, achieve men in order to get to work. And do you know a etra? And and that's actually a super interesting discussion. And I think the first step in that discussion is actually understanding the banking this hype up term A I, what that actually means, but also where we are currently.

I the the idea is that the technology itself is not evil. This is something that it's nuances. So it's not played up in media narratives or movies and novels like this, but it's very new as the not events, what we do with IT that determines what that does to us.

And so a good example of that is facial scanning. So you know, the country that I live in right now, a lot of people in the west are very scared that facial scanning is mostly in the hands of the government. And these cameras all over the country that are able to sand people's faces, almost twenty four, seven. The most .

common stories about china, right?

yeah. And you know, this is social score. All that i'm sure a lot of your listeners have have heard of that. And so, you know, that's horrific to a lot of people in the west, because no privacy is something that is held. Sacco, you know, another data point is that in china, crime has dropped significantly. And criminals who you perhaps have been on the lamp for the last few years, some surface because facial and software to a public safety in many ways has been vastly improved because of this.

And you know, if you think about the problems that china is facing, and give me an example that a lot of Young parents might relate to in china is a lot of issue around safety in in nurseries and in early education facilities like daycare centers, where, you know, there was a, there was a big incident, two thousand eighteen, where these Young kids were caught on camera being harassed by their teachers. Teachers were hitting them rubbing. I think he was mustard in their eyes, making them eat muster to punishment rib.

And there is nothing the parents could do until many days later when they found the children were um they were actually harmed physically. And so if you have technology that can, for example, sense anxiety within a room increasing, or that catching things on camera and actually filming these actions and then ferrying that this type of action is not standard to a classroom behavior, warning the parents immediately, warning, you know, the other teachers immediately, and perhaps teachers could intervene. This is where the technology can be used in, perhaps to help away.

Now, of course, you can also be abused in the new. Ask yourself, how much trust do we put into the people who are controlling this technology? If you come down on the line of, no, I don't want anybody controlling me, even if they end up doing this for my own good, then you will cast a more suspicious I towards A I, but know these things also.

They lead into conversations about government, the world government, power government. And so ultimately, I look at IT as a problem that we have to discuss together as humans, that I will will be want to live in A I is a tool that we now have now available to us. How do we want to deploy that tour? Can we come to an agreement across society, around those use cases? And can we actually take advantage of the benefits of these tools provide us? These are our questions that um I questions are human questions.

Yeah and so in some ways, actually elon musk has a really big point there.

You know should we let privately controlled companies by a very a few selected founders whose entire business idea is to, you know, make money to create this new technology? Or should we let government in countries with less transparency, control all of this? Or should this be actually some think we all own the control together? I think that's a like a really good question that thanks to the topic of A, I actually gets raised much more now and enable us to actually talk about. And you definitely have a point there that we need to separate the human problems versus the technology because at the end of the day, we, together as a society, decides on how we use that technology, what we use that on, and and also a how to help controlling, in many ways.

A I most like a reflection of us because IT analyzes human patterns and like recommendation engineer, suggest things that we would like. So we talk a lot about the the silo de nature social media. You get to see the things you I can, you get to interact with the people who are like you.

And so if you look at at that way, A I is a reflection of what we want IT to be. So when I, without getting too philosophical, these are the things that we have to think about. You know, if if we are a society that that really benefits from the additional conveniences of knowing exactly what I want when I want IT, then I can do that.

I mean, I think I really just another example of a time where I was um and this is powerful n in china, I was browsing something online is related to a certain product. And then when I went to when I went to walk through the mall near my house later that afternoon, a sales person came out and suggested that I I look at a product and I was I was a different problem, the same product as when I was browsing my mind. So the online in the house are completely integrated.

And one person I think, well, that's great, like, you know exactly what I was looking for, that remarkable IT says me a lot of time and feel like you you know me yes, that's nice, nice. But for me, I was weird that I was like, why you know me a too, really, and take a while to warm up to people and what else you know about me, you know, you also know type of drinks I like, and you gonna that as well. So is this what you want? And you is different because, like, a lot of these things are decided in the aggregate.

So you know, a company make decision when enough people want IT that I mean, everybody wants IT. That goes back to the point of can you turn this off? Can you shut off? The a tracking will come.

We allow you to do that. Are you even aware of what's being tracked? This is the whole issue with facebook. They on you and how much much do they give um these things, I think would benefit from being more transparent and giving more choice to the person, the end user. But if the end user is an knowledge doesn't care about that is ultimately A I is not going to make the decision to be quote unquote virtuous or um exploitative because these are human Normative terms. I doesn't know what is right and wrong .

and with that, thank you so much, nick, for join this special episode of digital china. And as usual, any feedback just reach out to us and any suggestions of new topics, please feel free to reach out. Thank you for listening.

Digital ally china is produced by me.

jack up when .

ivell and tom show, and is powered by radio, an independent media platform .

expLoring culture, innovation and life in china. You can find them at radio china dot com.

Thank you for listening.