What should we do about unbelievably intelligent AI? | Ben Ross | TEDxMacquarieUniversity

I can tell you how you're going to die odds are is from one of these ten things on this list of ten dementia and Alzheimer's comes in at number three dementia and Alzheimer's Alzheimer's is a disease which deteriorates your brain one in four Australians over the age of 85 have Alzheimer's every 67 seconds someone in the u.s. is diagnosed with Alzheimer's if you have a proclivity to have Alzheimer's your only chance is for technology innovation to come up with a cure because today out of that top ten list it's the only one without a way to slow it prevent it or stop it so this is poignant for me because my great auntie Susie she died of Alzheimer's just a month ago and before her my grandfather died of Alzheimer's a couple of years back so for me finding a cure for Alzheimer's in my lifetime is really important and for that I am cheering on advancements in technology I am relying on it and I'm sure that for each of you you have your own personal reasons why you would want technology to advance but Alzheimer's is mine so the good news is that technology is definitely advancing we know that but the bad news is that as technology advances it gives birth to a whole nother issue one that not many of us have given much thought to and that's the challenge that computers will become smarter than us humans we call that a human level of artificial intelligence and when computers are smarter than humans and can perform better than us in all sorts of ways we call that super intelligence and for folks that know a bit about super intelligence they call it the last invention we'll ever make and the last challenge we'll ever face Stephen Hawking said that so how likely is it that we're going to hit this level of human level in artificial intelligence it's a pretty high bar right like we're humans and our brains are pretty sophisticated well it's a factor of two main things advances in hardware and advances in software so let's have a look at how hardware is going so since the 60s a guy called Gordon Moore predicted that for the same price every 12 months computing power would double and that's held true for the last 50 years the impact of this law on us is that I have more computing power in my hand right now than this supercomputer of the 1980s called deep blue and it's not just hardware that's getting more powerful software is actually doing this double team where it's making the hardware even more powerful so software is advancing to a point where you can use a computer game today and this is an image of Kevin Spacey and it's hard to tell the difference between a computer animated version of Kevin Spacey and the real-life video image yet one's a game and so the advancements in software and hardware mean that the artificial intelligence that's available to me today in this mobile phone is fast advancing to become at a human level of artificial intelligence and the interesting thing here is that it's going to happen faster than we expect and why is that well as humans we think linearly quite easily so if you take 30 steps linearly you get to 30 1 2 3 4 but if you take 30 steps exponentially like doubling each step you get to over a billion and Moore's law is telling us that technology is doubling every 12 months so we're on this exponential curve so from this exponential curve when's that magic moment where computer intelligence overtakes human intelligence well you get a bunch of experts together and their average prediction is that that magic date is somewhere around 2040 so that's only 25 years from now and the point at which we might hit super intelligence will their average prediction for that is 2060 my view is that even if it takes twice as long it's going to happen in yours and my lifetime this is a thing for our generation and when the time comes that a computer is smarter than the human brain we're going to look back on the human brain as if it's some sort of low powered low memory retro computer system like this one and that's kind of hard to get your head around but if you think about it the human brain doesn't really stand that much of a chance because human brains trapped in our head it's stuck in a physical container and it's difficult to expand the size of it unlike with the computer you can add memory to it add computing power to it upgrade it you can easily reproduce a computer whereas many of your parents in the room will know the sort of investment you need to make to reproduce the human brain probably the most powerful piece that a human brain can't compete against is I think a concept called collective intelligence and that's when you string a bunch of computers together and a little bit of information that's somewhere in the corner of the computer network is able to be scanned and into the the knowledge base of the whole computer system and no matter how much we munch our heads together or sit close to each other we can't do that with our human brains it's difficult for me to read the mind of someone in the back row of this auditorium let me give you an example of how powerful this concept of collective intelligence is so those folks that were working on that supercomputer back in the 80s deep blue well pretty quickly after then they focus their energy on creating an artificial intelligent computer and so to test themselves four years ago they pitted that artificial intelligence against the best humans at a general knowledge game on TV and guess what the computer won so why did it win because it was able to process over 200 million pages worth of information including the full text of Wikipedia in less than three seconds now that's a pretty powerful thing to contend with right so so what sort of future might we have if we're contending against such powerful computers well it could be awesome so from my perspective I need a cure of Alzheimer's in my lifetime or else I'm likely to hit get hit by number three there right so for me it could be awesome super intelligence could create some sort of nanotechnology that goes into my bloodstream and clears up all those decayed cells in my brain maybe makes me smarter better-looking grow a little bit more hair whatever but some not just inside of my body like generationally and environmentally could make a massive difference for example super intelligence could create carbon scrubbing technology that cleans our air and environment from greenhouse gas emissions and carbon dioxide and solves environment big environmental issues as well as the top ten diseases so it could be awesome for us but it also could be apocalyptically bad so imagine if super intelligence decided that it didn't want or need to share this planet with us or it wanted to save us from ourselves and wipe us out so these are pretty bad outcomes and in the in the context of human history we haven't had many technology advances that have posed what's called an existential threat on us a threat that we could be wiped out but this is one of them and that's why Stephen Hawking says it could be the last invention that we'll ever make and the last challenge that we'll ever face but we're at this great point in time right now where a human level of artificial intelligence is still 25 years away so we can do something about it and my point of view is that folks like you and me can create guidelines and start to develop the pathway forward to make sure that we get to a great outcome rather than a bad outcome so what do I mean by guidelines well I was particularly inspired by Isaac Asimov who thought of a set of guidelines back in the 40s before there was even robots or artificial intelligence and he developed these three laws of robotics the first law was that a robot may not injure a human being or through inaction allow a human being to come to harm now that's a pretty nice little guideline right so let's think about an example that's even closer to close to us than the 1940s in the 70s there was advances in biotechnology for the first time we were able to recombine DNA and this was a pretty dangerous thing because we could have created all sorts of mutations and viruses and a bunch of folks just like you and me got together at a place called Asilomar and they develop guidelines for how to use biotechnology in a safe way and since then we've had all the benefits of the biotechnology revolution and none of the detriments and that proved to us that by the guidelines are a good place to start now I don't know about you but I feel a bit nervous that we're currently heading towards a human level of artificial intelligence and we don't have any guidelines in place and in fact each of us in our own way are actually helping to advance technology so for me I spent my days developing software apps and I enjoy that but all of all of us when we search the internet or we send off a tweet we're actually creating and curating the fact base for super intelligence so my point is that if we start to manage this we can make sure that it doesn't become unmanageable thank you you

Channel: TEDx Talks

Views: 97,557

Rating: 4.3808947 out of 5

Keywords: TEDxTalks, English, Australia, Technology, Computers, Disease, Health, Intelligence

Id: -KaGZhB0WkI

Length: 12min 40sec (760 seconds)

Published: Wed Dec 02 2015