Not If, But How Artificial Intelligence Might Take Over the World | Hugh Baillie | TEDxISKL

did you know that you most likely have AI in your pocket at this very moment did you know that I am the century change the face of this world the next century you probably didn't that's why I'm here to talk to you about today AI or artificial intelligence is defined by John McCarthy of Stanford University as the process of making intelligent machines specifically intelligent computer machines this definition does include your phone your laptop your computer this projector your calculator AI has become very integral to our society think about it has any of you heard of y2k y2k was a epochal apocalyptic prediction of um from 1999 P believes that once 1999 transferred into 2000's computers was shut down completely they couldn't handle this change of date and thus computers be wiped out and humans in the process does this tell you how vital AI is 12 to our survival and now a eyes come around so quickly and most people don't really realize it here's 1990 in 1990 any country would has less than 25 computers per 100 people even the most developed countries didn't have more than 50 computers per 100 people in 2006 these numbers skyrocketed countries had more than 19 computers or even 80 computers per 100 people and we're nearing a hundred computers per 100 people that's one computer per one person in a measly 16 years this is a huge development in terms of AI and the reason we're so dependent on it is pretty understandable on a macro level governments use this to collect data store data and organize their citizens because in China we have one point you know more than a billion people to manage that's a huge amount of people and without AI we could simply couldn't do this in America 300 million and a is coming so vital to our survival on a micro level I use computers to study you might excuse computers to work I need my calculator to do math and it's only becoming more more important and so what does this mean well it means that AI really controls us in ways you may not expect it's also replacing jobs for humans I'm sure most of you might have seen or used a self checkout station at a supermarket and what it does when you put a knife through it scans and gives you a tally and the price of your groceries and it doesn't require human and they're incredibly advantageous to supermarkets because you can fit a higher concentration of self checkout stations in a small space and they don't need feet they don't need food they don't sleep they don't human rights and they can work 24/7 without getting tired and so it's way better than human but they're still limited they can't react to new situations they can't clean the floor they can't answer questions that people need answering and so we still need humans to be at the supermarket but what if we didn't then that's the next level of AI it's called a GI or artificial general intelligence sorry artificial general intelligence is expected to be Delap developed by 2040 and will change the face of this earth so now think about it differently AGI as defined by the AG our society is um artificial intelligence that matches the speed of the human brain or just matches it and mimics the human brain this means that they can think like us they can speak like us they can be creative like us converse with us just like humans think javis I'll go back to that later in terms of supermarkets we could completely wipe humans from the supermarket because if a GI was there they could react to new situations they could learn new skills they could answer questions that people have that people need to be answered without having been asked and before been programmed to answer them this is absolutely huge and now what this means for us as a society is you might lose your job if you have a if you work at a supermarket or maybe you have a similar job and it might sound bad it might sound like huge unemployment but it could also mean specialization a long time ago made thousands hundreds of thousand years ago we started developing tools and started keeping animals and farms and this mean that people wouldn't have to hunt as often and we'd have these people who didn't have jobs and so you'd imagine that to be a bad thing all that means they could specialize in two new categories and actually accelerate human development and make it more efficient and this same thing would occur hundreds of thousands even millions of people wouldn't have a job but this doesn't mean that they they would die this means they would find a new job in a new field and just accelerate human development and so this is these are huge implications another example is the AI Butler Mark Zuckerberg CEO of Facebook has announced his plans to develop an AI Butler for his family similar to Jarvis this AI Butler could recognize faces could check on the baby and manage the household at all times very useful but still limited once again by what it's been programmed to do a G I on the other hand would changes completely AGI in a sense could recognize new faces and sense Li essentially store new data and think in a different manner for those of you have seen Iron Man Jarvis is kind of a witty funny weather look at a spy storm that really you know everyone loves Jarvis and when he young spoilers but when he died nearly dies in the Avengers everyone gets really scared because everybody loves Jarvis right and that's what a GI could do they could actually make jokes and think differently they could really change this earth and they could essentially replace humans but what's holding us back from a TR well first step is hardware I mentioned that the T on to a supercomputer bill in China just matches the speed of the human brain that same supercomputer also covers 720 square meters and cost half a billion dollars to build and takes a lot of power but our brain runs on this 20 watts and it's very small but in 2025 we will have the same processing unit for simply $1000 in the size of a regular chip this means AGI could exist on a much more a much grand scale and much more efficiently this is the first the next step on the other hand is software is the thinking process it can't just think just because it's fast enough it needs to be able to think like a human tim urban author way but why believes there are three methods to achieving AGI his the first possible method is simply replicating the human brain copy it like a math test you seen in class and this markets next to you and he's got a way better test than you you know he's going to do way better than you fact he's already done the test and he's already 100 percent might as well copy him that's what we're doing we're simply copying the brain but it's not so easy brains are tiny neurons and they're much more complex than we might imagine that's what the next step the next possible option is is replicating evolution it a whole ton of computers give them the same task whichever two ones do a better mash them together make them a bunch more times and we're simply doing evolution at a much faster rate the third option it's a bit more weird basically give AI the ability to code itself give it access to all information and artificial intelligence and let it run free kind of imagine giving a baby a room full of crayons and every artwork ever imagined and saying makes them are that's essentially what we're trying to do no one knows which one's going to get there first but people believe you should be able to reverse-engineer the brain by 2030 the real problem is that AGI it's cool but it's not that great in comparison to ASI asi or artificial superintelligence is down right frightening ASI exceeds intelligence of humans it can be thousands hundreds only double as small as humans but even a slight increment increase in a intelligence can really just change this earth think about last century what have we developed we've developed an MRI a nuclear bomb we developed much faster transport mechanisms overall just doubled our life span and change the human race forever in simply a century flat-screen TVs computers AI itself and so that's in a century now if we double the human intelligence that growth would be exponentially higher triple con temple ten times as fast and if we are making an ASI ten times as smart the possibilities are endless ASI in a sense could end it could create immortality or creates extinct it sounds very science fiction but it's actually a very real process Stephen Hawking I'm sure who many of you know um believes the AI could spell the end of the human race for very good reason because once a si is developed it's going to exceed us and it's going to have its own goals in mind because it's going to think independently and if its goals aren't allow aligned with our own it could be bad news real bad news but like I said it could be it could create some immortal it sounds totally crazy something out of a comic book but so we've doubled our lifespan in a century if we were twice as smart or five times as smart by now we could essentially end the deterioration of telomeres in our own bodies reverse the aging process cure cancer create new methods of environmental engineering and essentially solve problems that exist today every single problem depends on the development of a si because once it's developed things will skyrocket undred real damn quick so why is it scary oh it's scary because it's going to come around real quick once a G is developed a si is very near on the corner because the advantages to a egi is that it can self-replicate much faster than humans as humans we take eighteen years give or take maybe more to self-replicate or as most people call it making a baby the baby is generally a little smarter than us and that's what we hope for and we can make a lot of them and so it's actually it's it's a it's a wonderful experience but the problem is that it takes a really long time AG I won't have this problem the thing is that I can replicate within minutes within hours within days and essentially come ten times as smart within a matter of weeks and once it's exceeded us think the possibilities really are endless the time it takes for us to hit a hammer a nail together they could be out colonizing stars and turning colliding planets and turning stars in the fuel cells it's really crazy and so why am I telling you this why do you care or the reason you should know is because we need to be ready for this change because it's going to happen it will people need a GI because people will need to develop that's what happens whether you like it or not and if we're ready for change things will go quite smoothly think about your grandma your grandma probably refused to get a phone or oral computer for a very long time and everyone got really pissed off because they want to call her but she refuses to pick up a mobile phone if for example we do the same thing with AGI it could be a rough experience because it's not just the object at this point it's natural entity that we can interact with and in the past refusing change is caused a lot more problems xenophobia and racism have been a very real thing and the same thing could occur with AGI and the problems will just be created if we accept this change but also be wary of it we could really develop as a society really really well and just it would be a fantastic experience if we develop a CAS I with the right goals in mind like I said we could be immortal your grandma might be around forever my dad might be around forever everybody you know might be around for the rest of humanity we might be out exploring the galaxies it does sound like science fiction but again it's very real and on that note it's tough I know asi sounds pretty scary like immortality extinction it's a bit of a bit of a hang of the balance but it's better to be ready than to be scared thank you

Channel: TEDx Talks

Views: 126,714

Rating: 4.0807219 out of 5

Keywords: TEDxTalks, English, Malaysia, Technology, AI, Change, Future, Science

Id: mDMP2t6lUjo

Length: 13min 17sec (797 seconds)

Published: Mon Feb 22 2016