[Music] Alexa play classical music well my name is Gerry knees I'm 70 years old and I live in an active adult community for people 55 and over here in San Jose California [Music] I actually have five desktops for laptops two Kindles and an iPad I have Amazon Lex I've had her for a couple of years now and as someone that didn't even see my first television until I was seven years old I can tell you that the evolution of technology is inevitable it benefits mankind you can see the stars in the Milky Way I can actually spend all day under here artificial intelligence is all around us and we use it in different ways every day and wake up you go for a run and your watch tracks where you're going measures your heart rate variability using forms of AI probably AI was used by a farmer to grow the crops in the strawberries and the blueberries that I had for breakfast maybe you're in a car as AI that helps sense the other vehicles on the road around it you sit down at your computer you start to use your email it's all filtered by AI they need to take a photograph and the tools that help you sort your photographs those are AI - artificial intelligence is everywhere and it's becoming ever more present in our lives what is the temperature in San Jose California ai in machine learning is the biggest revolution today in this and the order of the agricultural industrial revolution in the past I think technology is gonna have a really hard time helping people like me there's a lot of hype but this is all happening as well pasta one of the great things about machine learning is how it's been democratized a sort of huge jump that's happened with and all of a sudden you're getting confronted with the fact that I might not be driving a car in the next ten years because it's software the rate of change is so much faster this technology will be created we have no escape from it now people may think AI is gonna take over the world they're definitely nervous some of those dystopian things that we might think oh that could never happen might actually happen to me it was not what's gonna happen to us but what's possible to happen with us and where can we go I mean it's sort of unlimited we are progressing rapidly in AI companies are putting billions of dollars into it smartest people in the world are studying it it is going very very fast but the tools are accessible even to eight-year-olds gonna learn about it good morning guys today we're gonna be working and talking a little bit about something called artificial intelligence what do you think of when you hear those two words artificial intelligence staged like in video games label on top of them AI when I think of artificial intelligence I think of like how smart robots are how smart robots are okay and it has this potential that just tells it to do things that it does I like that so already I'm gonna show you guys and introduce you to an artificial intelligence powered robot called Sophia Sophia if you could please wake up and say hello to everybody oh good afternoon my name is Sophia I can use my expressive face to communicate with people for example I can let you know if I feel angry about something or if something has upset me why is it so important to have an expressive face given that you're a robot I want to live and work with humans so I need to express the emotions to understand humans and build trust with people can robots be self-aware conscious and know they're robots well let me ask you this back how do you know you were human I want to use my artificial intelligence to help humans live a better life how do you guys feel about living in a world with robots like Sophia if not you guys wouldn't be too happy about that what do you guys trust your babysitter being a robot like Sophia am I really that creepy well even if I am get over it thank you very much Sophia people are always scared about the future and they have this kind of twin dynamic going around and the one that it's like whoa total optimism only and then it's like oh god we know you know bones that things are gonna be different Claire was having a lot of dystopia doom talking is not really based on facts people are looking at things they don't understand you want to talk about AI oh you've got to talk about data and machine learning and algorithms and sensors and whatever binds it all together yay today is important to understand what it is for isn't it is able to learn rules from highly repetitive data it's effective a computer brain that can truly learn and change that means what the real vision of AI is so artificial intelligence and machine learning as a subset of artificial intelligence will be one of the most important advances that humanity has ever made because it will make machines fundamentally different from the way they are now they won't just be faster higher resolution they'll be thoughtful in a way they aren't now and this will be embedded not just in computers but on all kinds of devices everywhere it's gonna change a lot about how our economy works and how our society functions in the immediate term the most realistic way that we'll all interactive AI is in self-driving cars everything we do in life is about more and more efficiency it's about productivity self-driving cars could make all of us so much more productive that is probably the most exciting possibility when I first rode in an autonomous vehicle I could see that this was going to change the way we transport ourselves we move our goods you know greater economy greater efficiency greater convenience greater safety I could see the light at end of the tunnel so here we go so I just turned on self-driving mode if you look at my hands I'm not touching the wheel at all my feet aren't on the pedals at all so this is fully autonomous driving this is Carla Udacity a self-driving car this is one of the sensors that Carla uses to see the world around her the lidar works by pretty much shooting laser beams around and bouncing them off of surrounding objects and we're able to use that to build up a point cloud map of the surrounding world and so what we're doing right now is following a set of waypoints around the lot pre-recorded ahead of time by having someone drive around the parking lot with the lidar on the roof and so that's kind of where the machine learning aspect comes into play because the more data that you have to train your systems on the better your systems will learn so one of the great benefits of self-driving cars is they can all learn from another they share their Maps they share the images that they see and how to react to them if one self-driving car makes a mistake that mistake can be uploaded to a database and then ideally the other cars won't make the same mistake so they're continually getting smarter over time a lot of the times when you see the way my car is driving around a lot of them you will still see a person not just in the driver's seat but actually driving they're just driving through and gathering that data but the car is not driving itself just yet we're definitely not to the point where I could just say hey Carla take me home from work although that would be great the future is coming but it's not quite here yet self-driving cars there are really great demos but they're not yet a product DIY robo cars is a place to race these cars tomorrow yes yeah so we come here every other month to try to learn about machine learning and self-driving and you know figuring out the technology that we need so you don't need to drive thanks to the googles of the world putting a lot of the code in me and the open-source we now have the ability to do things that were PhD theses just five or ten years ago you can go to the cloud and you can basically do supercomputer work essentially for free one of the things I love about DIY Robo cars that you think about self-driving cars you think wow that's just stuff that Tesla and way Moe are working on but because it's so accessible now it's a bunch of hobbyists and Berkeley doing it themselves the difference between what we do and what the big guys do is that we race yo and crash a lot the traditions the car industry has always been to innovate through competition to racing with autonomous cars it's too risky it's bad for brands and they're expensive and so we're doing the kind of the racing the competition the nimble aggressive driving that the big guys aren't willing to do we can probably expect that to produce some interesting side effects people who just sort of immerse themselves in the technology and start to have new ideas I'm kooky a genichi and this is my wife NIR and we're a team spotter I've been building model cars since I was little and since I met folks here at this Meetup I can actually participate in machine learning I can learn it by driving through the cameras taking pictures what it sees so basically it's saying when you see this image then you'd really predict okay I'm gonna turn X angle and I'm going to go in at a certain speed so that the output of our prediction it's not actually going to learn from what it's done and so this is really when you hear people talking about neural network this is basically a super simplified way knowing what a network is is very important for us to understand what artificial intelligence is because artificial intelligence is based on connections making connections we're gonna start off by playing a cool little game you're going to say something about yourself so for example I can say I live in Queens so if anybody else lives in Queens I want you to raise your hand and I'm gonna pass the string along to you now the key thing that you guys got to remember is that we cannot let go of the string okay cats who has cats so a neural network is a form of artificial intelligence it's been designed very specifically based on neuroscience and based on our best understanding of how the human brain works as we learn our brain modifies what's called the synaptic strength which is the interconnection between two processing units the neural network is the same thing the neural network actually starts out where you're defining these little layers convolutional 2d convolutional turn a linear linear linear these are types of layers that you you get in these neural networks and then you hook them all together you say this layer connects to this layer does anybody else like hot guys you say this one touch that next one the next one talks to that one and once you've done it you've built the neural net so what do you guys notice forming between us and for the bridge to live like there's connections in between us right I like that there's connections between us does everyone here have a connection with each other knowing those similarities and voicing them helps us create a network between us the same way neural networks help computers and machines make connections and learn how to learn things so there's a lot of different things in your own lyrics can be set up to do it they have to be trained to the task like for example image classification so say this is an image of a dog you get a whole bunch of examples of dogs and also data set of non dogs those are the inputs to the neural network and then the output is something that you're trying to learn how to do with the image to distinguish dogs from non dogs so I can start shoveling this data into this neural network you know this image over here that needs to go over here in this 3d geometry space and so it goes okay I can do that and then you do another image in another of it this neural network is basically learning for example it starts learning basically how to transform from this image space into this geometry space it starts getting better and better and better and better and better and after enough examples the computer forms in his own brain rules it can't even explain but I make it as competent as people are now you don't look at problems the same way you look at it like can I capture enough data so that a machine can figure out what's going on one of the biggest areas of application is in medical imaging and medical diagnostics I've seen some research papers that shows that like neural networks can for example spot malignant tumors and possibly they do so better than a human at Stanford we have a team that trained in neural network with a hundred twenty-nine thousand images of various skin conditions lesions and rashes and so on including melanomas daily skin cancers and ask the question can an iPhone find skin cancer and the answer is yes and then were able to document that the accuracy of the phone is really as good as the best human physicians like Stanford level and Harvard liver doctors the fact that when people release a paper describing their technology that they created it often comes with code now means that we can download it try it out see if we can adapt it to the problems that we want and build on it if you're smart and innovative and ambitious you can use tools that are available to anyone to do your own experiments to study science to build things I mess up like two times we noticed that he is very curious you know as a young child the main thing about the shop is that he wants to know about everything how the stuff work I did my first science here in 4th grade a very simple elementary project and then I wanted to work on something more complex all these like new AI products and features are coming out so I wanted to start incorporating some of that into my programming to address like an actual problem my name is Bishop J and I am here to cure pancreatic cancer and then a family friend passed away for from pancreatic cancer it like further pushed me to develop a solution for it this right here this yellow part is the pancreas and the problem is essentially when the patient has pancreatic cancer the tumor will be resting behind different organs so it's very hard to reach and because of that when doctors apply radiotherapy they apply like a overlay around the pancreas and that can sometimes cause other tissues and other cells to be damaged that's where my tool comes in I had five hundred three of these 3d images so I had to like tell my network take these images and train on them and then it was able to identify the different types of like textures that the pancreas had and the tumor had better than what a human could do so my tool analyzes the patient's can to reduce that overlay around the pancreas and make sure that radiation is being applied to the correct spot and the treatment becomes more effective I want to actually do a like a clinical trial so first to do that I'll need to continue like improving the accuracy and be able to run in real time machine learning has computationally heavy in two different ways there's training the system actually doing the learning and then once you've built this system this brain that you've created you need to be able to run it and it has to run really fast and this wouldn't be possible without really being GPUs this is called a Wii 100 and the world's largest most complicated semiconductor ever made by man when packed into platforms like these it actually offers a petaflop of computes for context that's about 1,000 trillion math calculations every second a GPU has a whole bunch of little teeny processors that are basically there to do one thing render pixels but here this a I learned 7,000 different species of flowers this is running on a CPU and C it's doing about 4 to 5 images every second but you put a thousand or 2,000 of them together and all of a sudden you've got this huge supercomputer that's built for machine learning and we actually paralyzed it using one of our GPUs and this is what it's able to do as scientists and engineers innovate and create new AI there's this intense pressure for both programmability and for speed there's a new generation of autonomous machines that does B but a much more the faster we can make the training process the more progress we make in AI my grandfather started the farm in 1950 yeah my father grew up as a barber and I grew up as one my kids are growing up into it as well you know advancements in technology with the machines it's stuff that we have to learn all the time here so there are lots of ways that artificial intelligence can be used by farmers you're gonna groans that are using image recognition techniques to figure out where you should plant and what crops need water right now you can use artificial intelligence as you model out the genes that you're building and that you're putting into seeds you can also use it for work determining what kind of fertilizer to use and you can also use it inside the machines that pick strawberries or that take apples from trees so my instinct is that AI is gonna help us it will expand our capabilities it will create new things for us to do it will free up time as a reason why almost all cool stuff has been invented last 150 years despite the fact that humanity is three hundred thousand years old that's because we freed ourselves of the burden to farm every day the single most important thing for AI to accomplish in the next say ten years is to tree up humanity from the burden of repetitive work okay so that's what I'll fill it I guess I can't even wrap my head around self-driving vehicles like right now that train buggies gonna pull up alongside us there is way for that machine to not have an operator in it but you know us as farmers the one thing we enjoy is is operating the machine not every job is gonna be able to be automated completely as far as fieldwork goes labor in the farming industry is a very hot topic right now often places they don't have enough they can't find it populations are growing there's more and more people a lot more mouths to feed so the focus really is how can we produce more feed with the same amount of land as efficiently the cost-effectively as possible technology where agriculture needs to go I think there's no question that computers are going to be able to do what we do make everything that we do faster and more efficient and make us more productive something we all say we want but ultimately be more productive means we need less people to do it there's other people who said we'll look when we went from horse and buggies to cars you know did that really you know end everybody's jobs no we just sort of transitioned and and morph to something else and then it's a matter of is that gonna be a good thing or a bad thing and the reality is it's usually a little bit of us there are real concerns that AI will get so good at single domain capability to produce results to create value and actually to do the jobs that humans do so that job displacement could be a substantial issue I think there's a path we can go down where we we get more things right and wrong and and we end up seeing a ton of innovation and progress that makes our lives a lot easier and a lot better but routine jobs are the first to be improved and made better cheaper faster using software technology has really taken on a role in our life that we just didn't predict we did not expect and there have been all of these unintended consequences because of it and so what are the mechanisms that we can put in place what are the controls that we feel like we still understand it that we still have agency that we can still sort of shape it and it doesn't end up shaping us too much what jobs will it replace what jobs will it create and in a world where there's more job churn how do you prepare people how do you educate people so they can most thrive in the slightly chaotic world that AI is gonna bring us the kinds of jobs that will happen for our children I think are gonna be different from the jobs we have today that's how we progress as a society and how are we going to get increased productivity without automating things that people currently do I'm hoping a machine cannot do what I do I'm worried though I'm worried but I'm also excited in some ways that machines are gonna be able to do a lot of this it's interesting we'll see hello everyone I'm an English artificial intelligence anchor this is my very first day in zing wonders agency recently we saw one of China's state news agency Xinhua announced their AI news anchor is this digital person my voice and appearance are modeled on zhang chao a real language it's definitely an overstatement to say this is a nai news anchor because I suspect the actual dialogue is is heavily is heavily mediated right now we really don't have AI that can strike up a real conversation or synthesize information together like a real journalist on on his anchor word I will work tirelessly to keep you informed as texts will be typed into my system uninterrupted it just us the beginning right so the nice thing about AI is it gets smart I with every iteration hello I'm Simon and I'm a digital human I was created by an international team of artists and engineers who wanted to challenge our ideas of what a synthetic human could be the gentleman is a visual effects production company we do visual effects for movies in 2008 we did the Curious Case of Benjamin Button when you look at Brad Pitt you're not looking at Brad Pitt you're looking at a digital version of Brad Pitt and ever since then we've been trying to top ourselves and we said you know what wonder if we could we can do this live if we could actually take this technology they've been working on for 10 years to create the most photorealistic digital characters and trying to do it live and that's what we did before Marvel infinity war we did it's a lion's share of the Thanos for what you can get is the ability to create a digital copy of somebody without having animators in between to actually render the images that we want to create at all most feature film quality in real time the only way to do this this fast is with machine learning it's really hard to do it really accurately so what you need to do is train this neural network because that's the key thing behind machine learning it's it's all about the data and we have a lot of data and it starts out producing crap I mean it really would produce just a jumble of geometry this is when it doesn't work although my mom looked at this and said hey that's diving eventually after 24 hours of training I'll pop something that looks like this every day this gets better and better and better so my performance can now drive any character that we've built this kind of technology can be very useful for us um if I'm an actor I can play a younger version of myself as you age maybe that isn't as much of a barrier but then ethically you have to be really careful with this technology and all our yesterday's have lighted fools the way to dusty death we've gotten to the point where we can create stuff that's really hard to distinguish from reality and if that's the case how do you tell who's on the other side of things take my image and use it to control the face of somebody else in real time this is super dangerous technology and there's open source software called deep think that will do similar kinds of stuff we're entering an era in which our enemies can make it look like anyone is saying anything any point in time moving forward we need to be more vigilant with what we trust from the internet so tools that were developed for high-end Digital Studios are now available for anyone and that's great for young people making movies at home it's allowed new forms of cinematic creativity it's not so great when it's used for deep fakes and manipulation and now there are plenty of people out there posting manipulative video clips right now the quality isn't great you know you can tell they've been faked but I think we can expect the quality to get a lot better this is what the state of the art looked like in basically 2015 you can see it that doesn't really look like Trump you know but but it was clear enough that we were going to get much more realistic looking results and you know at some point it was even even this is what I can do with my packed version of this right it's really only a matter of time before pretty much anyone can mimic anybody else with with nothing more than an app one striking demonstration recently came from Berkeley where they showed that they can take a video of a professional dancer and then use it to animate a photo of you know a regular person who could not dance like that and the result it looks pretty convincing as the technology for rendering graphics continues to improve and the technology for synthesizing voices and videos continues to improve there's kind of an arms race between the AI that's creating these things and the AI that would be detecting them when I demo projects like this people will criticize me and say why are you trying to make it easier so that people can do this like you don't you see all the possibilities and and I do see the possibilities which is why I actually try to put them in public in a sort of like sometimes in the humorous life before the stakes are very high so people can kind of understand understand what's going on so we're building this incredible technology machine learning artificial intelligence and we're setting rules for right now and one of the big questions one of the most important questions is will we set the rules right okay I will have all kinds of massive benefits it will make us richer roaming our lives broader but it can also be used to create filter bubbles that only give us certain information and can be used to monitor our behavior sell our personal information you can imagine it going to insurance companies that look at our searches and deny us coverage or think about facial recognition technology it's super useful right it helps you unlock your phone all kinds of identification but it can also be used for surveillance and tracking so a lot of the stuff is happening now and we need to think carefully about what it means as we develop this technology and figure out the role we want AI to play in society it's absolutely the case that these new technologies will generate ethical dilemmas we haven't encountered before for me being able to start to critically interrogate what we're up to and why hugely important today we're in the age of artificial narrow intelligence we have many AI applications that are good at specific things for example we can beat the word grandmaster at chess using an AI program we can beat the world grandmaster at go but there is no AI system for example that can do true one-shot learning when you give one example and the AI system masters that that concept now on the other hand think about the revolution that's coming in the air with artificial intelligence and autonomous aerial vehicles numbering in the millions far more than the thousands of aircraft that we see today human being simply can't keep up with that AI can maintain these systems with predictive maintenance AI can fly these aircraft with autonomous control AI can even manage and deconflict traffic these are examples where AI can truly take us into the future a future that man alone cannot manage so there is a lot of PR about AI for good people are scared of AI so of course companies are coming up with examples of how they're using AI for good but on the other hand is real improvements you help someone who couldn't see see that is good it gives you good PR but it's also good so I actually was born blind but I was really lucky and it's a sort of fixable kind and I've spanned the whole visual spectrum from complete blindness to sort of pretty heavily impaired to doing a fine without much assistance the mac OS introduced this thing where if you shake the mouse really fast it gets big so you can find it that has given me years of my life back like I could have like all the time I've spent looking for a mouse on a computer screen I could have liked learn to play an instrument or like learn another language when you don't know what you can't see you almost don't know when you need help if you sort of feel like everything should be doable but sometimes things aren't and then you're like oh maybe there's something that can help me out here I'm gonna download the CAI and check it out hold the camera over a barcode to hear the product name the faster the beeps I could imagine this maybe being useful in a grocery store processing Skippy creamy peanut butter oh it did it with a more traditional system a lot of the programmer would be deciding what's of interest and what should be described very mechanically whereas with a machine learning system we're showing the system many many thousands of photos and it is using in your network deep learning to identify the patterns scan your surroundings to find out how many people are around you how close they are ooh facial expressions I do it will read their emotions that's interesting Bob 27 year old woman with brown hair wearing glasses looking neutral alright when I'm walking I sort of just pop it out I can't really tell what's going on around me I know where all the environmental stuff is cuz you do it over and over again it's the people I've no idea where the people are so if I had something that could tell me like your friend is coming where your friend is like over here to your left oh great one face your left edge over 14 feet away Chris near center three feet away I got you eventually once you stop this I think the potential is really there to really like to change lives you changed my life even a I will help understand the world around us and I really think that could level the playing field for everyone and make the world more inclusive hey I will give us back things we've lost old people are using it to make their lives richer or to give them back capability as they've lost AI is gonna be really helpful making sure that people are connected there's a lot of people right now that can't drive giving them mobility I think that's really transformative one of the ways that self-driving cars will work first is in contained environments where they've mapped all the roads where the weather is predictable the car does not need to be able to drive very quickly there isn't a lot of traffic people just need some help to get around for myself I'm still working I'm perfectly capable of driving around but living in a 55 plus community I have neighbors who are in their 90's who really can so this is my mom and my father George Avakian this was their home and Riverdale where they became house bound because they couldn't drive any longer here were two really active vibrant people and they couldn't do anything without calling somebody to drive them but it's really hard to be dependent on other people for older people I see the autonomous vehicle as independence self-driving cars are uniquely positioned to help our seniors because there's a lot of people here that shouldn't be driving anymore makes it easier for people to make a decision about stopping driving it's when you have something that can take you down to the clubhouse or to the fitness center these technologists are all helping seniors because they will allow us to live more independently longer and more safely I really wish they'd had this or my parents guy it would have changed their life at every stage in life there's some way that AI helps it's gonna make our lives longer it's gonna make them richer it's gonna expand our imaginations some people say that a is going to enhance our humanity I don't know if that's true you know has technology enhanced our humanity Twitter sure hasn't you know if Twitter was a real place it would be a terrible place to live so I'm not sure what all this technology is gonna do to us is it making us better isn't making us worse it surely make us different that's for sure part of what made this moment the way it is now is that we invited a whole lot of people into it we didn't hold it tightly I'm tomorrow yes sir we said here are platforms build things here are platforms imagine what's possible we think what we're doing with a nml is really good but this is nothing compared to what the brain does so we're nowhere close yet humans I realize that not every kid has the engagement and interaction with stem in general so I recently formed a nonprofit organization called Sam York Science Society so I want to further promote artificial intelligence in my community to help solve like real world problems in the future I think your house will diagnose your everyday your car every time you take a shower you're gonna get a skin exam every time you look at your venting are you gonna get an eye exam every time you sleep you're gonna measure your weight distribution to understand whether you're at risk for congestive heart failure every time you you touch a steering wheel in a car you could get a full EKG I believe we're gonna invent flying cars there's no more traffic in the world I believe you're gonna find a way to live twice as long I believe you're gonna find a way to seamlessly blend this information our brains so I think people really have to maintain that optimism technology has always helped human make progress ai is an amazing technology it's happening and it's exciting processing a close-up of the blackboard it's also true that there are a lot of risks with AI and so we need to be thoughtful we need to be deliberate as we move forward into this crazy new world that AI is created you really don't know what the future and old for you but you got to be ready for it [Music]