The Terminator, SkyNet and Alexa: The Present and Future of A.I. | Marc Talluto | TEDxIWU

Video Statistics and Information

Video
Captions Word Cloud
Captions
simple question for everybody how many of you feel like you interact with artificial intelligence every single day all right about five how many of you feel like you are pumping information into the cloud every day that could be used to manipulate your life in some way Oh fascinating interesting okay so there's a combination there so two things I want to talk about or actually kick this off the fourth Industrial Revolution so this actually there's a quote from the current Google CEO sundar Pichai that we are in an AI first world the fourth Industrial Revolution I kind of see it as the sneaky breakthrough moment it's happening it has happened all the interactions as you'll see are actually heavily influenced by artificial intelligence so let's take a look first there's an old phrase we used to use about having myriad data and a dearth of information like I just I just have a whole bunch of data but I don't really know what it means but let's take a deeper look data is the fuel of machine learning how computers learn about what they're doing what they need to do next how they should act how they can win games let's take a deeper look first off make a separation between data information and knowledge so there could be an abundance of data but you really don't have any information if I was to say the stock price of IBM today is whatever 100 could you act on it you would say well I don't have enough information I just have data points so okay so data over time is going to give you information on what you should do but the knowledge is gained from information over an abundance of time in certain situations now you're actually thinking about other possibilities outside of just the stock market think about the data we actually produce so there's two different things to think about one is the data of things weather data all the time temperatures humidity a little dissolve xn slater sensors there are sensors on everything anybody here have a Fitbit okay has anybody heard the news recently of fitbit's being used to report on army locations and movements around the world okay sensors like Fitbit they're everywhere we are sensor izing everything smart readers smart meters the Internet of Things are all reporting data about our lives smart fridges smart appliances our utilities all of these devices even engines on machines our cars they have system logs they can all produce data and you've ever been shopping on something on Amazon and then you went to your Gmail and there's an ad for it right we're not quite creeped out about it anymore right if this was 15 years ago whoa how do you know what I'm shopping for now I go yeah I'm going to buy that but all of these things are talking in the background let's talk about you two so now your age figure out your gender might know your weight your relationship status you give that up freely your income you give to the federal government I know by the census data people know and can find out and just say no the federal government gives up income tax data not enough but anonymously to data engines to figure out what zip code what marital status how much money you make so guess what people can target market you and will keep going every time you like something every picture you take and share every time you click through something every movie you watch or pause or rewind everything you purchase which credit card you use or if you use PayPal is being recorded and tracked so people can figure out what to market to you what to sell to you to make other influential decisions and we willingly give all this information up what's really important is this aggregation over time when you put it in context you have results did I win did I lose this is a fascinating fact did you know that one there's a higher amount of smog the stock market goes down it's true it's been proven small correlations and how much fine particulates are in the air around Wall Street's affects behavior and so on those days stock traders aren't as risk-taking so the stock market declines a little bit just a little bit but it is noticeable and there is correlation so this creates information information that is actually extremely useful I have to make this comment because our cybernetic overlords are coming and they might already be controlling us but well we'll get into that but there's massive acceleration in not just the computing power but the ability for the computing power we have to discern through all of these piles of information we've been producing in order to make meaningful decisions for us big data if you haven't heard of this term it's kind of what I was mentioning before we have stockpiles of data but we don't necessarily know how they're correlated but we're saving it all we're storing it all does anybody know the NSA is storing every communication every cell phone conversation every email you have in a two billion dollar facility in Utah and I swear I don't believe in aliens I'm just telling you it's actually happening in the world they are they're storing all this information so we have these pools and data data that we sometimes we don't know if it's relevant but we know we have it feeding this is the Internet of Things if you've heard of this every device smart televisions your phones of course we are all basically giving up our location data or giving up what we wear we're driving how fast we're going you can argue that your device is an Internet of Things for people but the Internet of Things also applies to like I mentioned smart meters cars we just have essentially all these sensors reporting on what's happening in the world now robotics is an applied application to using this information in more the physical world artificial intelligence is what I'm here to talk about this is taking all of that data allowing computers to actually program themselves to learn things and we don't necessarily artificial intelligence the difference between artificial intelligence and machine learning is that we don't know necessarily what the results will be we program the computer look for things to make correlations so try to figure out something we don't actually know we don't point it to go play checkers we pointed an information set and say and say tell me why stock market dip so I should even say stock market dips and they say you know we found out there were smog on that day things like that that's where artificial intelligence is going and it's even going farther so what's good let's take a look a moment and think about it every does everybody not use Netflix or Hulu or something else okay let's think about how you used to go to the movies you used to buy a generic ticket and the theater didn't know who you were that went about what the movie theater is what the studios knew was that somebody bought a ticket to see a particular movie somewhere in the world at maybe at that particular movie theater that was kind of it and we've seen movies bomb we've seen you know there's been bad movies what Netflix knows about you okay sorry who you are I know what you watch I know when you watch it I know when you start a series I know if you finish a series I know what you like them didn't like because we put three stars five stars we actually put thumbs up thumbs down I know when you now Netflix knows there are five people in your family I can tell when you log in and watch a series but you're your companion you chose to watch it without them at some point I can see that this is actually happening so what does Netflix done with this information has anybody seen house of cards stranger things they've gotten into original content they don't make things that bomb why because they already know what you like they already know what you're interested in and they know who to market it to I know your age I know your location I know how many people are in your family I know that my kids watch way too much on Saturday morning that is the difference that's how this information starting to be used so this is still the good so let's move forward so automated farming vertical farming I know how much water I've given a plant cameras on these plants I know when they're actually growing I know what the proper temperature is I'm using all these sensors kyoto japan currently has a plant that's producing ten million ten million heads of lettuce in a given year through vertical farming using artificial intelligence driverless vehicles some people that argue dry texting means it is a driverless vehicle trust me I do it myself but actually it is this is real it is happening okay so the car manufacturers are doing a GM Ford Google Apple uber lyft cities across America these cars are driving presently Tesla and not only is Tesla doing automated driving they're sharing the results of driving and where to turn and what they did that one car what it does Tesla actually reports back to the mothership and that makes improvements to its software so all the other cars suddenly learn from what you drove and how you drove so the actual auto driving mechanism and Tesla is self learning and it's adaptive day by day ah the medical field nothing is more awash in data about you your weight your blood pressure you know certain diseases may be genetic diseases other things you may have but like for MRIs scanning MRIs scanning one if you wanted one doctor to scan one our Mir MRI what would you like to see a million MRIs scanned and try to find the pattern between them so they can determine if yours is at risk for a disease currently computers can has a 90% accuracy on diagnosing skin cancer there is no doctor that is that good this is the one that is pretty impressive sorry alphago alphago if you're familiar with chess and checkers like I mentioned go is a game a little bit more sophisticated more strategic so Google said they said that could never be beaten by a computer human couldn't be beaten by a computer not only did alphago beat a human actually two out of three within life he was in the last two years they then built alphago zero which actually did more deep learning more artificial intelligence to train itself it then beat alphago a hundred times straight then they pointed alphago zero at a chess basically just how to play chess they just gave it instructions within eight hours it beats it beat the world's best chess computer so that sounds positive that seems fun right the other side of this is the flash crash now this was eight years ago high-frequency trading algorithms if you're not familiar with this stock the stock market is not controlled by humans saying I want to buy stock they are controlled by algorithms that are programmed to use artificial intelligence to say mmm it's there's weather here this is gonna affect butter futures I see information about the Nikkei index how is that going to affect the stock price the problem is those bots actually trade against each other I actually came out of the high-frequency trading business years ago they actually put simulated fake trades to see what your bot is gonna do against my bot now my trains are battling it this flash crash was actually the Dow lost 10% of market value for about 12 minutes back in 2010 they say 60% of all trades on stock markets are actually done by high-frequency trading algorithms I wouldn't want to be killed by a robot but in the Geneva where they have the Geneva Convention just last month autonomous lethal robotic systems is a big concern we're familiar with drones but the actual button that says I should drop this bomb it still has human intervention they're now trying to prevent those going to the mind of a computer say well I see an intruder I should shoot this person it's kind of like Robocop but this is this is very real this is both with air warfare missiles and ground-based systems there is if you're familiar with Elon Musk who runs Tesla Bill Gates with Microsoft Stephen Hawking the famous physicist they've all been warning extensively against a robotic future of warfare the strange thing is you know who leads in the development of autonomous weapon systems the United States do you know who then is following suit because we've led the charge China Russia high-confidence Russia interfered whether or not there's collusion is not the question here the ability for a foreign nation to buy ads to make sure what you click on and what you like influences what you see and how you perceive to then make you act and do something that maybe you wouldn't have done is an effect of artificial intelligence so if you've seen Facebook in the news is changing their newsfeed they're trying to they're trying to dial back actually all that automated news that you're receiving I don't know this is kind of leads me into the weird there is an AI religion I kid you not it's funny it's actually WTF but I don't know what to do about that but the way of the future basically saying that intelligence shouldn't be just limited to a biological intelligence will we not have a higher power that has all this information all this calculation power isn't that a higher intelligence that we then almost become robots for so now the weird I can't explain some of this stuff but the best restaurant in London by TripAdvisor didn't even exist that's right some guy thought it would just be funny to create an address create a phone number put an ad online and he game the system people would call him to make reservations he'd say now sorry we're full sorry we're full I believe people gave good reviews to it and kept pumping it up pumping it up it was the number one reviewed restaurant in London it was his literal shed and it did not exist so you may heard the story of Alexa dialing 9-1-1 so there's a domestic domestic violence abuse and somehow the cops just showed up and they said well Alexa wasn't able to make a voice call like well somebody called 9-1-1 and we can actually hear the arguments so somebody in this home dialed 9-1-1 so it's like that Alexa just intervene no tell you the Microsoft AI that they put out there to communicate as a chatbox became a racist and a sexist within eight hours and they had to shut it off language creation so they had AI BOTS actually talk to each other and chat to each other they started talking in their own language they were still using English but then they started to use their own manipulation of English to actually make it faster to talk to each other Elon Musk as I mentioned before is actually a big he's very much against the the future of AI because he's very much concerned what the impact could be the humanity so he's actually built a company called neural Inc they're just forming to help build brain computer interfaces for humans so he can take all this interface and we could take all this information and actually use that information ourselves how I'm not too sure there is the IAI creative arts there is music here's where it gets weird if you get run over by a autonomous car if you get shot by a autonomous gun who's really to blame is it the company that built them is it the programmer like you know the entire legal system is based on somebody being accountable for that action now it's actually artificial intelligence one thing about the fourth Industrial Revolution is that AI has been said to increase GDP by 17 trillion but the work forces of advanced economies they rely on those financial job like I mentioned that are going to be easily actually easier it easily replaced and this is a 2030 estimate so we're talking only in 12 years you have high skill high pay jobs actually that will still be promoted people that actually be coding these things maybes more sophisticated type of work that can't be replaced by AI but really low skill low paid jobs this disparity will continue to increase we always talk about the wealth gap and people who recovered from the last recession this is just going to escalate even more and more the upside is there will be we talk about a massive elimination of waste in the supply chain to know who needs what and when and where the logistics behind this is becoming far more sophisticated I'll have to say this too this is not something we're going to stop it isn't the benefit of every company to do this to eliminate this waste there is really no stopping the progress that we're making in artificial intelligence just the thought think about what we do to try to prevent nuclear weapons from giving out of hand if you had a proliferation of artificial intelligence based weapons that are actually smaller and easier to smuggle but some despot in some country could easily you know suppress people with it probably a higher proliferation of those weapons one thing I will say is China is betting bigger than we are so the impact that on the US China will actually have a double their GDP impact because they're making a bigger bet on artificial intelligence so let's think about this for a second AI tools are available for everybody to use Google has and Facebook azzam you can go out there open-source you can go on the cloud right now and start using them these aren't mystery programs that people you know hiding somewhere and why because they want your data the more you use it the more they'll collect your data so if my choices are already being selected and filtered the news I choose to read the routes I choose to drive the movies that are being chosen for me if I'm being influenced I often get fake data potentially like that restaurant if my decisions are being based on what I read and experience my education my news can be faked right but it is being selective for me i artificial intelligence sometimes even written by healy effects the food it's actually feeding me those farms the Logistics I'm kept alive with that early detection to make sure I live nice and long but at some point I could be given false data my Apple map or actually don't use Apple Maps my Google map could drive me off into a bad part of town I could be shot by an AI robot am I already living in an artificial intelligence world that's all I got so thank you
Info
Channel: TEDx Talks
Views: 250,221
Rating: 4.6859159 out of 5
Keywords: TEDxTalks, English, Technology, Computers, Data, Robots
Id: QZ_NJkpdG2A
Channel Id: undefined
Length: 18min 22sec (1102 seconds)
Published: Wed Feb 28 2018
Reddit Comments
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.