 Ken Aizawa, he's the Vice President of Engineering at the Tech. Welcome, I was about to say, Ken, you know Caribou. I do know Caribou. Finally, welcome home, your home, literally. Thank you very much. Thank you very much. All right, so Ken, what has been your best experience so far here in Kenya, for how long have you been here? Yeah, so I've been in Kenya for about a week now. I would say the best experience was definitely going on safari in Masaimara. And but also just being in Nairobi, I think, and getting an experience of a city really embedded into nature. I come from New York, I was born in New York, I live there now and work there now. We don't really have such diverse animals, plants, trees, forestry, and all of that kind of mixed in with these high-rise buildings and bustling streets and all of this. So it's a really, really nice change. And I would love for that to be a part of New York, just more green. Oh wow, have you tried any of our delicacies? I've tried, is it called machama? Machuma? Machama, there we go, that was last night. And I think I did pretty well. I think I did pretty well. Yeah. Okay, so Ken, for someone meeting you for the first time, who is Ken, where is he from? From New York, of course, and what do you do actually? Yeah, sure. So I was born and raised and in around New York. I still live there and work there now. I studied computer science and philosophy in school. In uni. And now I work more on the computer science side, as a first as an engineer, now more as a kind of engineering leader in the organization. And I work at a company called The Take, as you introduced. And we work on AI technology for TV and movies. And my focus has been on developing that AI, now managing a team that also that develops that AI with me and deploying that AI across millions of TVs around the globe. What is your most treasured memory while growing up? My most treasured memory growing up, that's a great question. I mean, I guess I would say there were a lot of really great learning opportunities that I had. One of my most treasured one was when I studied music in New York. And yeah, I spent some time, I played the clarinet, which is like a musical instrument. And yeah, and I studied at the school called the Juilliard School in New York. And yeah, that was a really wonderful time for me just to be completely focused in music. Fantastic. So starting this conversation when it comes to artificial intelligence and just what the future looks like. What is artificial intelligence and how does AI work? Yeah, it's a really great question. It's a big question. So I'll start by saying that I think AI is sort of a misunderstood term. I think we read a lot of books or see a lot of movies. We see robots, killer robots or some program that talks and speaks like a human. But I think it's important to know that it's not zero and then one. I think there's a lot in between. And in a sense, we already have artificial intelligence today. The way we organize companies and we have certain people doing some things, other people doing other things, coordinating across organizations, across the globe, across supply chains. And we use software as a part of that. We use machinery as a part of that. We use computers as a part of that. That whole layer is a kind of intelligence. That's not one human being, right? It's many, many human beings creating this kind of new thing, a company, a government, a state. But artificial intelligence is a technology. And I think what you're probably referring to is software specifically that is able to process large amounts of data and turn that into really useful actions. And what I'll say about that, though, is that that is also already here. When you are on Amazon shopping for products and you are getting a recommendation for something, that's AI Outwork. It's saying, oh, he or she bought this one product. And so I'll recommend this other one. Also on Google. Also on Google. Or when you're on Netflix, where I spend too much time on. But anyway, if you're on Netflix and you're watching a piece of content and you like that piece of content, Netflix is saying, okay, this person like this content, what would they like? And let me look at everyone else that's like this piece of content and what they liked. And the final thing that I'll say about artificial intelligence is that because it's so multifaceted, because as we already have it, the gains and improvements of artificial intelligence will not be straightforward. It won't be one to one. It might automate certain jobs in the future. Like I think truck driving is one that'll happen in the US at least within the next five to 10 years. It might even automate some of the work that doctors and lawyers do. Right? So you both on all sides of the spectrum, both high paying, low paying, middle paying jobs will be affected by artificial intelligence. Are we heading into a world where we will completely build intelligence that can be able to mimic human base? For instance, the point you've said, like most of the careers will be tempered or like if it's in let's say an industry where this production going on, machines will be involved. So that means more people will in the future might lose their jobs. So I will head into that in the future whereby artificial intelligence that we build it, that we all completely just destroy job opportunities and we just focus on the creating aspect. Yeah, I think that's a great question. And I'll start by saying that I can't predict the future, but what I genuinely believe is that artificial intelligence and the technology behind it is only able to automate very repetitive and kind of time consuming, but very simple, otherwise simple tasks. So data lookup, data entry or driving or things like this. And there are jobs that will probably not be able to be automated ever, like creative jobs, for example. Right? Because we have to remember we're doing all of this, creating all of this automation, this prosperity to serve people. And AI is feeding data? Yeah, exactly. Okay. And as a result, AI will not be able to do things that only people can do, like creating art, right? Or in the hospitality industry, working with people and being with people, right? So I think what I hope is the future and what I also believe is likely the future is that there will be what I call friction in the transition. There's always friction in massive change like that. And it might actually be somewhat of a painful kind of change if there are 3 million truck drivers in the US that might lose their jobs with this technology. But over the course of 50 years, I hope that brings humanity towards a place where people are doing jobs that are better suited, more creative, more human, and to serve human interests better. Are there regulations put in place to know how far is too far? And is it actually regulated? I can only speak for the United States. I will say that I don't believe, I think there's some legislation for regulation that's currently in the works. And I'm really glad you asked this question because it's a very big issue in the US, I think, and a growing one. The regulation has mainly revolved around use of facial recognition technology. And there is a scientist, I guess leader, CEO of this organization called the Algorithmic Justice League, Joy Bualamwini, who has been really shedding light on some racial bias in some of the very commonly used facial recognition algorithms in the States by companies like Amazon or Microsoft or Google. Essentially what she found was that people with darker skin color would have lower accuracy rates for facial recognition and would be misidentified more by those algorithms, by Amazon than people with white skin. And that becomes a problem if police departments in the United States are using that algorithm to track suspects and try to identify criminals. That's true. So it's a huge problem and I don't believe there's been any significant step forward in regulation yet, but I'm hopeful that it will come. What are some of the strikes being made as AI is concerned by the tech? In the context of our country we have AI's credited scoring, it's using companies in accounting in terms of algorithms and models to determine whether someone is qualified for a loan. So what are a couple of strikes that the tech has made, courtesy of AI? Yeah, so we have a pretty fun use case, thankfully. All we're interested in, to put it simply, is to identify the people, products and places inside movies and TV shows. And to surface that information on the TV or maybe through the set top box or online if you're watching online as you're watching so that you can learn more about your content, so you can interact with your content. So let's say you're watching your favorite reality TV show and you like the dress that some contestant or character cast member is wearing. We make it possible for you to buy that on your TV. We offer other options as well that are a bit cheaper or more expensive depending on what your kind of price range is, but we use AI to identify all of those products and enable that on hundreds, thousands of hours of content. How do you guys do that? So I mean the nitty-gritty of it is simply we have lots of data, lots of images, and we tell the AI learn from these images and we train it on massive machines over hundreds of hours and the AI slowly teaches itself, learns on its own how to understand those images, the video, the actual frames from the movie or the TV show, again across hundreds of hours of TV shows so that I can do that automatically. So can I let you to help us understand something here? So let's look at Google for instance. You can search something today or probably next week then after a couple of days later you come back and just try how to it reminds you or it gives you feedback of what you did previously or even another instance which is a common scenario if you search about anything if it's a product later on you'll find an ad on Facebook or Instagram how does that happen? So that is the use of data by these advertising companies uh essentially they're tracking based on your you know Google username or email address or your Facebook username or Instagram handle they're tracking what you've clicked when you've clicked in what order who you've talked to and they are taking that information and saying okay let me use some I don't know if they're for some of these examples if they need to use AI but I'm sure in some examples that they really they are let me make a prediction based off of this behavior online activity okay and let me show some advertisements it's it's interesting I won't go into too much but some of the work that I do which is also related to advertising is kind of in opposition to that kind of form of advertising which really uses user data a lot and uses kind of this understanding of user behavior a lot to power this advertising solution rather than that you know our company we focus on looking at the content so what is the user looking at are they watching a movie about weddings or are they watching a tv show about dating then we can use that or are they watching a sports game and then so we look at the sports game not what the user is doing online on facebook or instagram or google whatever we look at just the sports game and we say how can we provide some advertising some useful information to the user just from that without looking at their data respecting their privacy and all of that yeah because if you're like in your busy studying somewhere else about something else google is literally just studying about you yeah exactly and I think you know maybe it's not the biggest problem but one of the concerns is that users are not very fit because it's on you still you know even after 20 years users are not very familiar with how google is using their data and where their data is going and so users can't really make an informed decision about whether to opt in or to opt out so and it's actually a free it's a free site so that's how they make their money through advertisement exactly but if google is making you know a thousand dollars per user per year it might be free which is great but maybe they could be returning some of that money to the users as well oh yes we'll be so happy yeah I mean I think so I think there's there's a lot of play there but it's a really really interesting question that's currently uh working itself out in the ad tech industry all right so we have telecommunication company that gives the ability for voice recognition uh is that also a branch of AI and how does that actually work yeah so the face recognition uh is an AI and the way it works is you have some what we call neural network which looks at an image of a face and it turns that face into a string of numbers so maybe 16 numbers and the numbers are such that that are produced from that face are such that a similar looking face right will have similar numbers okay so we compare these numbers with each other because the AI has digested this face to have some understanding of that face and we compare those numbers to see are these the same face or not and we do that on a large scale over hundreds millions of faces and that allows us to have a really nuanced understanding of of a face of variations in I don't know the shape of a nose or your eyes or your mouth or your lips or whatever but importantly we don't teach the AI any of this stuff manually it learns it all by itself okay is the same thing with the voice recognition as well exactly yes okay so most people confuse their three-term knowlogies these AI then there's machine learning and data science what's the difference yeah I mean to be honestly everyone kind of confuses that yes and part of the confusion is that no one really knows what AI is right because some people because some people think AI is some robot that speaks and acts like a human other people think AI is just a facial recognition algorithm so I won't comment on the AI piece I will say generally speaking data science is simply the science of analyzing data using statistics statistical models so you look at a large batch of data let's say it's consumer shopping data and from that you would do an analysis and you say okay what are the most popular categories of products and where should we place products on the shelves in the store that's data science it's analyzing a large amount of data and applying a specific statistical method like let's say linear regression or something like that to math data machine learning is an algorithm or is a group and family of algorithms that learn from data so they are actually trying to produce specific outputs from data and AI is probably at the top of the machine learning family all right so how is the artificial intelligence transforming like the media industry yeah I mean it's it's interesting because the media industry at least in the states is very slow to move maybe and hopefully that's different here it's very slow to move I would say our our company our product is probably one of the more advanced of the companies that are they're using AI to transform the the media industry and there is another way in which AI is used which I've started to see lately which we can use AI now to create images and video realistic looking images and video and so there is recently a show on disney plus it's a star wars show and there was a character from many many decades ago that they want to be in this new episode this new season but obviously that character that actor doesn't look like that now so they use some other actors some other young looking actor and they use AI to transform this younger this other actor's face to look like this other a completely different actor's face from 20 years ago the cloning exactly and they also used to that to transform his voice from his older voice to a younger voice and that all all of that technology is AI so I think we're going to start to see more and more content that's just created using AI all right so as you wind up like to find out what do you like to do outside work do your leisure time yeah I mean I like to go for a good run along with the broken waterfront in New York okay but otherwise I'm really really interested in art so going to art galleries and and museums and visiting those kinds of things in New York as well and then yeah and travel I would say oh and that's why I'm here fantastic we you're much welcome any other day this is home now right yes say this is home happening you buddy there you go you got it yes so people how can people reach out to you do you have social media handles I think the best way is on twitter okay ken a zwa all right thank you very much Ken as you are creating time to be with us today looking at taking us through artificial intelligence and helping us understand how the future would look like thank you very much I appreciate you having me all right so guys make sure you stay tuned you have another interview coming your way so we'll be right back