Watch videos with subtitles in your language, upload your videos, create your own subtitles! Click here to learn more on "how to Dotsub"

Algorithms Are Taking Over The World- Christopher Steiner at TEDxOrangeCoast

0 (0 Likes / 0 Dislikes)
There are people who will tell you... Crazy people People like authors who tend to be prone to hyperbole that algorithms are taking over the world But perhaps, we should take this thought and examine it for just a few minutes here We already know, of course, that algorithms have taken over a Wall Street. They make 70% of the trades that control your IRAs your 401 K's, your pensions. They controll your money, all of it. Our stock markets have become one layer upon layer upon layer of algorithms to the point where discerning order from the chaos for a human is next impossible Sometimes that's impossible for the machines as well That's what happened on May 6, 2010, what we now call a Flash Crash when a trillion dollars disappeared in 5 minutes Or in August 1 this year when Knight Capital lost 440 million dollars in 45 minutes when one of its algorithms went berserk I have a friend who is a fund manager. And two years ago we're out to lunch and I said to him "Is it true that algorithms are taking over finance?" And he said "Of couse!" But the interesting thing he said was "they're taking over everything". Now interestingly enough as algorithmic science advanced on Wall Street People began taking these methods and peeling off and taking them to other places and starting many revolutions in all sorts of fields medicine, sports, music... When we talk about these things, these algorithms... We often talk about them like we know them like they're our buddies If someone ask you "How that new Web app works?", you'll be like "Oh, that's just an algorithm" Well, what exactly is an algorithm? An algorithm, quite simply, is a set of instructions written in computer language that informs the machine of what to do with the piece of information Algorithms take input and they produce output Often freshmen and sophomore engineering students on their first computer science courses have to write an algorithm to play Tick-Tack-Toe That algorithm's input are the moves of the human, the output are the moves of the computer. But algorithms have gone far pass this now They can learn, they can adapt, they can evolve. They've evolved to the point, in fact where we are not always shaping the algorithms - they are shaping us They shape our culture, they shape what we see, they shape what we hear, they shape how we live I think we all imagine there is some line where algorithms can't get pass. They can't do those most human of tasks, right? But actually, these places are now too the province of the bots What do I mean by the most human of tasks? I mean things like grading students' written essays creating original art, making crucial national security decisions, writing legal documents Music industry already employs algorithms to find new artists They're very good at finding pop songs, because they know the math behind the best pop hooks They're already deciding more and more of what plays on the radios We have to ask ourselves: "Would the algorithms find Nirvana?" Would they find the Beatles? Your doctor will some day be an algorithm We already have a fully robotic pharmacy running on algorithms at the University of California, San Francisco It's told out 2 million prescriptions without making a single mistake Average human pharmacists would have made 20 thousands mistakes in filling out the same prescriptions You will meet an algorithm in the emergency room someday The question, of course, is this bad? That answer has yet to be determined. The story of the next 20 years is a story of Big Data and Algorithms We are the at the giant fork in the ark of humanity Just how much we allow algorithms to take over? But the better question for today, the better question for right now is... how much have they already taken over? One of my favourite examples of how far the algorithms could go and how far they have already gone is about algorithms that perform accurate psychological evalutaions on people This is something we normally reserve for humans who've been at medical school for 10 years How can a bot that doesn't know of itself, know us? This story starts 50 years ago, on the late 1960s when NASA made a decision to begin sending scientists in the outer space Up until this point astronauts have been plucked from the Airforce They were test pilots. These are the men who flew the planes that broke the speed of sound that rode next to the stratosphere, that spied on the Soviet Union These men were unflappable which is why they made such a good astronauts. NASA knew they would not crack under pressure. But there's still test pilot program for scientists How would NASA know which scientist would stand up to the pressure and which scientist would fall? How would they know which personalities would clash after being locked in space capsule for 72 hours? The Russians, NASA knew, had more than one mission compromised because of a crew conflict So NASA set out to create a system of personality classiciation a predictive system that would foresee conflicts between people and know who would perform well and who would crumble in outer space During the next 20 years NASA created this system of the most advanced psychoanalytic methods in the world They knew what you were thinking, what your personality was and what your capacity for stress was after a 10-minute conversation It can take humans years to master these methods But the important thing about these methods, the reason we're talking about them today is they're quantifiable So, how did they work? Works of the words that we speak, the way in which we structure sentences the way in which you use pronouns and verbs All these little things offer clues to our inner personality and how we work with other people In words of course, just like the momentum behind our stock markets or the math behind our music, words of data That data can be organized, stored, parsed So it figures that somebody would come alone and take NASA's incredible science and marry it with the new algorithmic technologies that can employ it everywhere So where do we run into these bots? We run into them nearly everyday And how do they know what we're thinking? Well, when the bots know your personality, they know the meaning behind your words For instance, my personality is something called "Thoughts based" When a thoughts based person is being explained... when somebody is explaining something to a thoughts based person They might sometimes say "That's interesting, that's really interesting" What I really mean when I say that, unfortunately, is "That's not interesting. I don't know what I think about this any more information" Now, a bots know exactly what a thoughts based person means when he says "That's interesting" So why do we run into these bots that know us so well? Well, I think we've all heard this refrain, a customer service refrain "This call may be recorded or monitored for quality assurance purposes" We assume that it means that once in a while a boss is listening to the call Then, maybe, yeah, maybe that's what happens. But often what happens is you've just invited 6 million algorithms in for a listen. NASA science was transfered to the set of algorithms So how does it work? Well, you call up customer service, the algorithms take note of your phone number then they route you to a random customer service agent, then they settle in for a listen They listen very carefully. Within 2 minutes have you started speaking, they assign you a personality They're incredibly accurate. The next time you call, they route you to an agent with the exact same personality as yourself and they can do this because a lot of these call centers have 15 to 20 thousand people on decade any moment And what happens when you get people with the same personality? Well, your calls are half as long. And they come to happy resolutions 90% of a time instead of 47% of a time. The company behind all of this has listened to 1 billion conversations. It has spent 60 million dollars over 10 years to create this library of 6 million algorithms to categorize the human language to build this mind reading bot. There is a warehouse in Minessota in the size of a football field where all these conversations, all this data is stored waiting till we recall that one file that matches perfectly your brain and how you talk Now we can do this with data and algorithms and we know who you are, why you called and what you're thinking It doesn't seem to be much outside of our reach The question isn't whether we will sort people this way. The question is where and how often? Will we sort job applicants with these methods? Will we sort college applicants with these methods? Or we even sort potential spouses with this? I don't want that person, you know... Will we sort children? Just with an any tool there's limits to the utility in all of this. We've seen the end of utility of automation on Wall Street It's become a place where humans have little insight as to what's going on and little control The people in charge of Wall Street have had a difficult time drawing a line between utility and menace As we go forward, data scientists, programmers and QAs in all sorts of fields will face the same dillema where to draw the line between utility and menace The story of the next 20 years is the story of Big Data and Algorithms That story will be determined by where these lines get drawn and who gets to draw them I hope you found that interesting, very interesting. Thank you!

Video Details

Duration: 11 minutes and 15 seconds
Country:
Language: English
License: Dotsub - Standard License
Genre: None
Views: 108
Posted by: dkalpakchi on Apr 5, 2015

Algorithms Are Taking Over The World- Christopher Steiner at TEDxOrangeCoast

Caption and Translate

    Sign In/Register for Dotsub to translate this video.