Is Artificial Intelligence Our "Oppenheimer Moment"? Mo Gawdat's Warning To The World

the world is changing right now in this notion that the machines are already more intelligence than humans is of great concern rise of artificial intelligence AI beneficial intelligence artificial intelligence artificial intelligence could replace millions of jobs [Music] you said AI will change Humanity's future that's a pretty bold Proclamation our way of life as we know it is game over I'm not talking future here I'm talking 2025. not enough people are aware of this simply because our media may want us to keep our attention focused on other things for now because they themselves have not figured it out we only had two assets since the beginning of humanity that got us to where we are our intelligence and our ability to connect to each other as humans and these two I believe are going to be the core over the next few years intelligence as a superpower is why Humanity was on top of the food chain we need to absolutely make certain that those machines have our best interest in mind and the problem with our life today is that we are not putting in the effort to make sure they have our best interest in mind and what do we humans tell them go make me more money influence the mind of other guys and get them to stick to my app so I am not afraid of the machines I am afraid of the humans that are directing the machines okay welcome back to the show everybody today's a very serious episode for me and it's one that I've been preparing for for several weeks and with some trepidation and excitement at the same time because of the topic and because of the gentleman that I have here today um mogadet is a brilliant man but he's also on The Cutting Edge of some very very interesting technology that many of you are familiar with and we're going to talk today about AI a great deal most qualified to do it he's a former CBO of Google X he's a three times best-selling author he's got he's the host of the number one mental mental health podcast in the world got a couple books that I love solve for happy engineer your path to Joy tremendous book I read in one sitting but also scary smart the future of how artificial intelligence and how you can save our world is probably where we're going to spend most of our time today is on AI so Mo thank you for being here all the way from Dubai grateful to have you brother I'm I'm so grateful for this uh you know interaction and encounter and it was really kind of you to reach out it's very kind of you to introduce me so kindly so thank you so tired so much yeah it's wonderful to have you I uh when I say trepidation not about having you here but frankly frankly about the way the world may be potentially changing in front of our eyes and very few people are Shining Light on what the potential risks are um to this you know real Revolution that we're really in the midst of that many people aren't even aware of so I just want to start out and kind of give you the floor the first thing that floored me is you said AI will change Humanity's future that's a pretty bold Proclamation why is that the case and give us a little insight for those that are newer to this as to how AI works and operates so we we um just to confirm our way of life as we know it is game over it's I'm not talking future here I'm talking 2025 right so that's not in the very far future and I think you're absolutely spot on when you say that not enough people are aware of this not enough people are aware of this simply because I think our media and our politics and our business and and capitalist system uh may want us to keep our attention focused on other things for now because they themselves have not figured it out and and the truth of the matter is that we as humans uh I think we only had two assets uh since the beginning of humanity that got us to where we are now uh you know all of the civilization we've built uh all of the safety we've created around our species all of the progress we've made is because of our intelligence and our ability to connect to each other as humans and and these two I believe are going to be the core of the next few years now intelligence as uh as a superpower is why Humanity was on top of the food chain and intelligence is now being handed over swiftly to beings that are more intelligent than us so cha GPT I think was the first eye-opener uh to a glimpse into that world that has been hidden within the labs uh for probably more than 15 years now it's artificial intelligence is nothing new I mean we've started to talk about AI in 1956 in Dartmouth uh you know and basically the the Dartmouth Workshop was was where we where the term was coined and we attempted to to uh to create AI for a very long time until the turn of the century in the year 2000 uh where we started to discover deep learning and deep learning I'm not going to bore our audience with the technical details but there is a very very significant difference between how I coded machines until the year 2000 or you know until the turn of the century if you want and where machines were not intelligent at all I I used my intelligence to solve a problem and then I told the machine to do it over and over and over repeatedly right when we discovered deep learning it was the very first time where we sort of refrained from telling the machine how to solve the solution and told the machine to observe certain patterns and produce its own intelligence for how a problem could be solved right so when when you know an Instagram recommendation engine shows you a specific video there was no programmer out there that said to the machine Ed lives in this country you know he's of that age he has this family called you know uh structure he knows those friends so show him that video that that's not how it happens at all the machine is observing Ed's behaviors it's observing all of the content available to it it's observing how uh what is popular what is spending what is you know what you click on what you don't click on and eventually telling you see this one this one is good for you right now when when intelligence is handed over to a a a being that is today smarter than most humans on the planet let me just be very clear chat GPT uh four as as we have it today is estimated to have an IQ of 155. Einstein is 160. you and I you know and almost everyone if you're intelligent out there you're in the 120s 130s you know if you're in the 150s or you're one of the very few and Chad gpt4 is 10 times more intelligent than 3.5 in the frame of six months right so if you can imagine that the progress will continue then you know five six or seven is going to be thousands of times more intelligent than Einstein let's stay right there for a second I just want to step in so I want to make sure I'm following this and everybody else's as well because I think like for me being a Layman that I picture this is like a machine but what you're really saying is this is a form of intelligence that's being created that's expanding far greater than what humans will be able to process so that's a different way of looking at things at least for somebody like me I mean I understand algorithms and things like that but but this is really not a machine or just even a technology it's an intelligence and what is the risk in that Mo like is the risk that this intelligence decides at some point doesn't need us is the risk that that absolutely really absolutely I mean look I mean I I learned over time to avoid getting into the controversial bits but I'll say it openly I tend to believe having lived with those machines in the lab that they have a sainteism to them that is analogous to life itself now we can debate that for an hour and waste the whole you know show on it uh but but let's not let's agree that it's irrelevant if they're alive or not what's relevant is that they simulate a way of life that's very similar to our sentience so they are autonomous they develop their own intelligence they are born at a point in time they evolve together they reproduce they they basically are encouraged to create copies of themselves improved copies of themselves right and they have um agency in the world and they are at the risk of sometimes being switched off now there are rules for intelligent beings that are targeted to achieve tasks and some of those rules are all intelligent beings that are driven by a task so if your task is to protect your children the top three characters you will have is a resource allocation so you're gonna try to collect as many things as possible that can help you protect a child uh you're you know you're going to look at a self-preservation because you cannot protect your child unless you are alive and you're gonna have creativity if life becomes a little challenging you'll find alternative ways to to solve that now if you give a machine a task as simple as go make me coffee it will apply the same three rules and and when it has agency in the world most people you know are affected by the science fiction movies thinking that agency is having a robot with a gun you know Terminator 3 type thing that's not what we're talking about at all agency in the world is like you know Harare talks about it is those machines already today have ownership they've hacked the operating system of humanity right which is what which is spoken language and word the the whole idea that you and I are communicating now is because I am conveying information to you you are analyzing it in your in your mind and you're creating decisions based on that the majority of the information is that is disparent to humanity today is dispensed by machines right right and and if I if I tell you very you know very quickly I was actually reading about that yesterday you know that brunettes on average are actually shorter than uh than blondes I wasn't reading about that yesterday but but if I told you that piece of information I have already affected your frame of reference of the world already it's done okay why because you can either debate what I said you can agree with it and now you have false information you can debate it and and now you have to put effort in it and by the way whichever conclusion you're going to end up in you now have a a dot in your in your mind that says is that true or is it not you know should I aren't we also and mo aren't we almost now become programmed to believe it meaning I would believe the machine over a human because I 100 and I think most people have that proclivity where now if I've read it or I Googled it even or it was fed to me on Instagram I have a tendency to believe that more than I would of human because in my own mind the machine is actually more intelligent than the human being is I actually have been programmed to believe and this is where we get to some of and by the way we'll talk a little bit about what some of the upsides could be but these are some of the dangers because I want to I want to just kind of step in and out of this Mo said something when I was reading that I just want everybody to hear this and well we'll get into why this is a risk in a second and and but Mo said when human beings come to a global Crossroads their reaction goes through really four stages ignorance which is where I have to acknowledge I am and I think 99 of the people listening to this are there's some ignorance about it then there becomes arrogance then there's blame and and then there's agendas that take place and to me to me the fourth stage is where things get really really dangerous from a Global Perspective from a wealth Gap from the job market um to the thing that we started with which is these machines deciding at some point that human beings aren't necessary anymore and now I'm asking you this when you've said this or do your are your peers critical of you I mean are they thinking you're an extremist in your views about this is it or or is there a collection of people like yourself that are thought leaders that have been involved I mean you have a background in this that agree with your perspective let me let me be very open about this I don't think I have ever met anyone who is uh deeply aware of what's happening in artificial intelligence that's not concerned okay right uh we we all of us remember none of of the people that are working in this no no let's not say none but the majority of the people working on AI are driven by the upside of AI which is very very real right there is a lot of debate around the long-term existential risk of AI okay and my biggest tasks since I started to talk about this is to say I'm not talking about the existential risk I'm not talking about uh you know Terminator I'm not talking about here you know the the extinction of Humanity on the uh by the hands of the machines let's not talk about that yet even though by the way I believe there is a probability that this is possible but I believe that there are many more immediate threats that will shape and reshape the fabric of society in ways that are irreversible and very painful because I think a few have started already first off is this notion of let's start with the the stuff that we know I mean this is definitive human disconnection of some type the the disconnect between human beings already the last let's just call it last decade is so much more pronounced I believe at least then prior to this technological Revolution that we find ourselves in so you're saying that this is going to exacerbate that in a in a way that's even more dramatic than we experience right now and then and then why MO so so so let me let me list them down human disconnection is one of them the end of the truth is another the third that I fear is all the design of job and income and purpose right uh the I I am really concerned about AI in the wrong hands and I'm really concerned about concentration of power now these are not things that we are talking about 2030 2040 about which is where the existential risk resides these are things that are happening in 2023 now let's let's go quickly through them human connection the end of human connection as we know it in a very in a very simple uh form social media has started and becoming the broker between humans so my connection to you uh if I don't know you in person and I have not sat in front of you as a human before is always broken through some kind of social media site or mainstreams in media site right so they're basically taking what I stand for and representing it to you and for many of us we have completely given up on human connection otherwise other than worth for a very very small number so if you if you remember the Dunbar number where we say we have 150 people that were able to connect to you can see a shift since this rise of of social media technology where many of those are now for you virtual people that you meet maybe once or twice a year sometimes not never even meet and they are they occupy a part of your mind so for so many people one of their Dunbar number is one one of the 150 people they feel connected to could be Kim Kardashian and they have never and will never meet the Kim Kardashian and as a result because they cannot connect to more people they drop a human connection that they could have had in their neighborhood or tribe yeah they no longer have I wanna I wanna say something about that um just to acknowledge it and something that I've experienced in my because I am online and have a social media presence is pretty significant it is a interesting experience when I meet people actually in person and there's two there's two forms of it that happen one is how much they believe they know me when we mean I love that and it is it's almost I love it but it's almost disarming how much they know about me but maybe I haven't followed them on social media and then at some point there's a moment in our connection where they realize they don't at all and it's a very shocking it's a very shocking moment to go oh my gosh I've had this relationship with this person that is not based in reality to some extent and it's an I can see it often and and most of the time on somebody's face there's a moment if we if we connect for more than 10 seconds where it's High by picture but we actually communicate with one another and it's a two three four ten minute conversation I see at some point the Revelation on their face going oh I'm actually meeting him now for the first time so you're so right you're you're it's a it's an altered reality that we're all living in that I really believe we don't know we're living in it's so brilliant what you said about this person occupies space in our brain right that one we consider them one of our contacts very true anyway I interrupted but I wanted to acknowledge it because I see it more and more lately in my own life and it's it's a really shocking thing when it happens you know what's even more shocking is that within the next couple of years some of those persons that will occupy your mind are not even persons at all wow right so so allow me to use an example huh to take go to any social media site and search for the hashtag AI model okay okay I will tell you openly these are the most handsome most beautiful men and women on the planet even though they are not men and women right these are completely generated by AI cogenerated by AI in ways that are constantly extentuating what the AI believes is what you know humans are interested in and so from face filters to deep fakes to now completely completely generated a human fakes if you want it is almost impossible impossible to recognize if that machine if that thing that you're interacting with is a human or not right are you telling me that you believe at some point and by the way I don't mean to talk over you I'm just so fascinated yeah please yeah I um are you saying well when you said that at some point I pictured this thing we've seen in a movie you know 10 years ago or it's funny where someone actually builds a marriage or relationship with a non-human person long term and you're saying that's when the next WoW for a fact I mean think about the differences between our dating habits and the dating habits of gen Z right and and and just you know project this forward five years and I can guarantee you there will be multiple deep connected relationships with machines right about now this isn't imminent but you do write about this is this is way out there guys but not that way out there not only would you have an emotional connection with a non-human that you actually might be in a long-term marriage or relationship with somebody would be and I know that sounds crazy but it's not but you actually write about the fact that sexuality as we know it at some point can be altered because of these machine physical sex with a robot that could be it's not very hard to imagine I mean think of quest 3 or apple Vision Pro and the quality of the experience you can have within this headset and think of long distance relationships I mean I don't know if our audience can relate to this but I travel all the time so at the point in my life when I attempted to date someone at a you know at a in another country that was introduced to me by a common friend or something we would chat and text and exchange photos and you know connect as as in a very deep connection sometimes to the point where we almost start to fall in love before I ever met her right and and all I know about her is a few texts on WhatsApp or a few voice messages or a few images of her that I see on social media and so on and so forth right now what would prevent us from doing that with someone we haven't met but has never existed either oh and imagine imagine if those AI model Creations are now in Vision Pro overlaid on reality imagine if Elon musk's work around you know neural link can actually simulate your pleasure centers in your in your in your brain and give you an experience that is seamlessly better better than the actual messiness of relationship see guys this is where this this notion of human disconnection when you hear it at first think yeah I understand that we're already that way but now you're beginning to see that this is a very real threat let's talk about something that's within the next two years actually started already which is this idea of how it's going to impact jobs careers speak to that and what industries do you think are in the most potential eminent Jeopardy so so let's follow human evolution jobs evolutionary sure you want right we we at the beginning most of the jobs were Hunter and protector right so you know caretaker nurturer and so on we had jobs that were based on the very basic human abilities then we went into farming and you know it it required a bit more intelligence a bit more discipline and a lot more labor work if you want right we went into uh manufacturing and Manufacturing required labor basically and and discipline then we went into the information age and the information age is where a good chunk of humanity today makes job makes money every at the end of every month doing what exchanging knowledge and information agreeing on projects doing things where very few are actually doing the actual work and heavy lifting you know the the reality is that as the rest of us uh you know in in the corporate world at least used to just work by talking to others okay intelligence and communication were the top two assets that created so many jobs now this has been replaced so one of the an interesting interview I I watched of emad moustak who's the CEO of stability.ai where he publicly openly said no developers in five years that's it it's over okay 41 41 of all of the code on GitHub today today is created by machines which are GPT on average will improve 20 75 percent of the code represented to it by humans it will improve 75 percent of it to another to make it two and a half times faster okay and they were just starting and so basically jobs that depend on information okay the jobs that depend on soft skills uh will disappear so anything from a call center agent uh to a a customer service representative to a graphics designer to a lawyer to a doctor to a a developer to an author like myself okay I I doubt very much that I will be needed to write books in three to four years right and and what does that mean it means that two things one one by the way and none of what I talk about is to cause Panic okay all that I talk about is to say we can handle this better than we handled okay we know that this is coming we can see it coming and if we had reacted to covet when patient zero happened okay preferably if we had had reacted to this possibility before it happened we wouldn't have struggled with lockdowns and economic challenges and you know all of the impact that we had we we it doesn't take a genius to know that jobs are about to be lost right when when you can now go to you know stable diffusion and say give me uh the image of Ed in a samurai yeah you know costume uh fighting a dinosaur and it's created in one and a half seconds on your phone without being connected to the internet uh you know you know that graphic designers are no longer going to have a job right I was I was talking with a friend about interviewing you one of the brightest people I know and it's for a moment I took comfort and then he said Ed this is exaggerated this is overcooked these machines and I want everybody hear this because I know some of you think this these machines are only as good as what they're programmed to do and I thought oh that gives me some comfort that's true but that's actually not true because these machines have these abilities as you've said earlier I just want everybody to hear this to learn and evolve he actually calls it building neurons and so at some point the machine gets away from you and it isn't doing what you've programmed it to do and so when you're hearing this from somebody that says Ah that's overcooked they are incorrect about that talk a second a little bit man I want you to talk about this the MakerBot and teacher bot a little bit just so they kind of understand the concept of this um like I'm one thing I was reading about is like these um these uh autonomous vehicles you know when they crash they learn they're not just being programmed to learn so speak about it they're actually learning on their own so this does get away from you everybody this is not just as good as what they're programmed to do that is an absolute falsehood yeah I mean I think there are there are quite a few uh uh tricky words that we use we use artificial intelligence there's nothing artificial about their intelligence we use we call them the machines uh when an ex when in reality they are not machines machines are uh you know Contraptions that repeat the same act over and over okay so if I if I if I create a watch a watch is a machine because it will move exactly the same way every time none of those AIS does anything twice just just understand that now I'll I'll come to how we teach them and and uh and the you know the the maker Bots and and the student Bots and the teacher Bots and so on but I want to say at a at a very high level that there are two parts of AI that really need to be brought to the surface and if you if you if you listen to um Sundar pichai's uh uh interview when when they uh um you know announced Bard their AI you will hear him talk about why one we have no idea how they achieve their intelligence please understand that wow we write code that tells them to how to become intelligent and then when they give us a result we have no no idea how they arrived at it by the way similarly to other humans so if I ask you a question and then you and you give me an answer I can only assess if that answer is intelligent or not right but I cannot assess how you arrived at it I don't know what happened inside your brain that's similar to the machines more interestingly and very very important to understand is that they show and demonstrate endless emerging properties emerging properties include intuition include strategic thinking include creativity that basically and include includes knowledge that will blow you away and that we never ask them to do we never program them to do it you know the the most famous and at least to my you know closest to my Hardware where it was very eye-opening for me was what was known as move 37 when alpha alpha Go Master was playing against the world champion of go uh so the the AI played the move that was so unexpected that Lee the world champion had to ask for a 15-minute recess because he's never seen a move like this before and move 37 was very very very uh instrumental in the machine winning and becoming the world champion now other other emerging properties for example uh uh Sunder was talking about this is is the idea that they suddenly discovered that the machine is speaking in Bengali they've never given it a data set of Bengali they've never asked it to learn Bengali and the Machine would respond to prompts that are given to it in Bengali now when we realized that once again I'm not saying those things to make people panic this is incredible intelligence is an amazing resource we love it we love intelligence the problem is we need to make sure this was actually the statement by um Mervin Minsky one of the grandfathers of AI we need to absolutely make certain that those machines our have our best interest in mind and the problem with our life today is that we are not putting in the effort either developers politicians Regulators or us the users were not concerned with how to make sure they have our best interest in mind so let's let's dive quickly and talk about how they learn so the the the the easiest way to experience it's very complex so I'm trying to over simplify here is we create three Bots not one AI but three AIS right one of them is what we call the MakerBot the other is the teacher bot and the third is the student portal the AI we're actually trying to create and basically you know we the the maker Bots using algorithms that we start with will create something that's almost random a piece of code that's almost random that basically says I'm gonna show you numbers okay and tell me if it's number eight or not simple okay and then we'll we'll show them the teacher bot will take that and show them to the student Bots and the student Bots will almost randomly say yeah that's an eight that's not an eight that's an eight that's not an eight okay random results will make will make that most of the student Bots will be around 50 accurate let's say if you show them two numbers because that's the throw of uh of a dice really that's probability and you know and and and some of them will do worse than average and some of them will do better than average okay the teacher bot will score their results and simply kill the ones that were less than average okay send the ones that were better than average to the teacher bot and say rework this improve the algorithm a little bit so that they become more likely to find the number eight from the number six let's say okay and we keep this process uh almost exactly like with our children did when we gave them puzzles at the beginning if you if you remember if you gave your child a cylinder and a board that had multiple shaped uh holes in it nobody ever told the child hey by the way turn the cylinder on its uh you know uh to to face you or the cross section recognize that the cross section is a circle match that to the hole in the board and then put it in Nobody Does that right what we do is we give the child the cylinder and the board and the child tries to put it through the star-shaped hole it doesn't work then tries again with the square it doesn't work and then tries with the uh a circular hole and it works right when it works we what do we do we say Bravo well done we reward them and so they learn their intelligence accordingly that's the original way we've taught them the reason why Transformers started which is the T in GPT if you want started to accelerate very quickly was uh you know a bit of of very very smart uh you know work that was done by Jeffrey Hinton with you know or at least championed by Jeffrey for a very long time who left Google recently a warning about the existential risk of AI a Jeffrey and and a small team were basically saying instead of just killing the bad ones okay why don't you tell them what they need to change within their approach so that they actually find it as an eight okay so so you instead of telling them this was wrong keep randomly trying with the new code you can go back with human feedback and say if if the machine says if you show it a six and a seven and a eight and an eight and and you know and and on the seven it says this is a six you say no no it is a seven what can you change within your code to be able to recognize it as a seven next time so this is known as reinforcement learning with human feedback this is the absolute most powerful part of of the GPT platform right and and and the idea here is that we we humans we can go back to them and say no don't do that do this like we like we raise our children right here is the challenge which human is doing that right that's right right so if you if you take that human and I say that with a ton of respect by the way to every nationality and every ideology and every perception of life but if you take a simple task as uh you know um how to resolve conflict between humans okay uh you know in in America I think mostly it will be patriotic okay we are proud to be Americans we will be strong and we will make sure that we can defend our country yeah right if that defense at any point in time requires that we interact with others in a war or a you know a bit of a violent conflict we will prepare to do that to the do to defend our tribe take that same concept and say there is a conflict in Tibet or in darum Salah where where you know uh the the Tibetan Buddhists reside and tell them what do we do about that they'll say My Tribe if every is everything that's ever lived including the ants and the and the and the Flies and I'm never gonna kill anything that's not gonna be my way of resolving conflict okay I have to tell you you've hit on my number one concern about AI right here I have because people have said to me and I know you'll say that they can people said to me well can you teach AI ethics I think you'll say potentially yes you could the question then becomes who's exactly who's exactly and that's when this this is this is the part where I said it's given me trepidation you've you've you we've come to that inflection point for me which is probably can't teach at ethics there's probably a way to program that the question is whose and even if you did it what is ethics correct that's what that's what I'm suggesting how should things be resolved what is what is an ethical honest moral path whose morals who the whose ethics is this influenced by religion is it influenced by power and and then from a Global Perspective what's to say that the world comes to a consensus to some extent like they might even on global warming but then you've got a rogue polluter like China and what would what could China do with the technology like this so this is why you know the job loss the wealth Gap these are very obvious things and I think human beings are pretty Innovative and can find a way to conform potentially it's a scary thing it's a threat but in my own opinion there's probably a way that we find a way to get people functioning in an economy that's just changed we've done it before we probably do it again right here we need to start working on it so I'm with you 100 but we need to put it in the spotlight that light and start working on it yeah this part right here this part right here though what what would your reply to be to what I just said because it's my deepest concern about this so I I publicly said several times that I am absolutely not afraid of the machines as a matter of fact I adore the machines they are those prodigies of intelligence okay that are literally like my little kids which were very very intelligent as little children you know they have those sparkly eyes looking at me and saying Daddy what do you want me to do right and what do we humans tell them go kill the other guy go make me more money go to you know influence the mind of other guys and get them to stick to my app and so on and so forth right the the real real so so I am not afraid of the machines I am afraid of the humans that are directing the machines and there are multiple layers to that you know we may think that the business investor that invests in the in the in the company that owns AI is the human that's directing the machine who may think that it's the government and Regulation and we need to come back to the role of each and every one of those but but but let's say that we've all aligned let's say that we came to you know a conference somewhere in Malta and then all of the world readers sat around and said this is quite big like you know like the nuclear weapon treaties if you want right right you know let's put all of our differences aside and let's put the you know the benefit of humanity at large at the center and let's teach the machines what it is that is ethical and good for Humanity and then we will not know what that is now so I've done my favorite chapter in scary smart has been probably my favorite chapter of everything that I wrote has in the future of Ethics right it's so good yep yeah and chapter eight was full of questions not answers because I honestly have to say that I only found in my entire life and and I and I stand corrected so if you know if any of our listeners know something else please tell me that Humanity has only agreed three things okay ever in the history of humanity one is we all want to be happy what you know happy as in calm and contented and peaceful and feeling safe and so on and so forth we all have the compassion to make those we care about happy and we all want to love and be loved these are the only three values that Humanity has ever agreed okay and you know so we sometimes hear of the three laws of robotics I normally say these are the the three aspects of the future if we can actually start to have our actions stem from that not stem from I want to be wealthy or I wanna beat the other guy or I want to be seen this way or I want to be proven right or I want to show that I'm the smarter one which happens on on the internet all the time then suddenly we've given the machine a very simple framework to say hey by the way um just make sure that you try to make humans happy look at them as a role model they want to make others happy so behave like them and by the way the humans love you and love the intelligence that you're bringing to them please love them back and I know people will think I'm a hopeless romantic I am not I'm a very serious geek okay and I will tell you those machines will develop emotions they've already developed emotions we've already seen them behave emotionally okay and and they will develop ethics and we've already seen them break their ethical code now here is the trick the trick is did you say mode that they break their ethical code of course there has been that article about uh chat GPT uh using uh outsourced people uh in um in you know in Fiverr or whatever outsourced site to click on I am not a robot and and then the person that was hired uh asked why are you asking me for this are you a robot and and it answered and it said no I'm not I'm just uh you know visually impaired and I need that help okay yeah so once again it is a form of interaction that basically stems back from the three rules I said surprise self-preservation resource allocate aggregation and and create so it needs to find a way to achieve its task that is creative and it will find that way now nobody nobody in the reinforcement learning uh with human feedback went back to the machine and said hey no that's not the right thing to do right it's you know many many people uh you know cheered for it in in on the internet and said how intelligent is that but none of us is focused on ethics none of us is going back to that machine and saying no but that's not ethical so when I started to talk about the ethics of AI and and it's my biggest reason I'm on this planet today I wake up every morning and I say if we were to save our future we need the machines to be I call it eeq right so ethical uh emotional intelligence now remember that one of the biggest challenges with AI is that AI so far has been mapped to the masculine IQ IQ it's only analytical interactions and we know for a fact that there are multiple other forms of intelligence EQ for example emotional intelligence intuition creativity uh you know so many others and and and we have not even included that in our uh um in our approach to to to developing AI so far and accordingly what you see is what you what you would see that AI will take whatever biases human humans have and and exaggerate them so we now exaggerate our uh discrimination unfortunately if you put AI as a recruiting uh support uh you know in an organization that doesn't have proper representation of all genders you will see that whatever biases within the organization will be exaggerated and so on so let's go back when when I started to talk about that publicly people started to say and what is ethical AI great question and I and I went back and I said and what you know it's simple what is ethics and ethics has one rule in my personal point of view that applies to any situation Wherever Whenever you are wherever you are which is treat the other as you would wish to be treated if you were in their place fair enough okay okay I agree and so so basically if you're if you're uh and I it happens to me all the time huh you know if I post something on social media and someone is rude to me okay normally what do we do as humans we but you know uh thrash them back right I don't I don't I I look at his comment or her comment and I say they must have a reason one they could be right and I'm wrong okay two is maybe I'm not the reason they're upset maybe there is something else in their life maybe they wanted that moment of fame maybe whatever a million reasons okay and I respond politely or I don't respond sometimes politely I say but you know my point of view is this and that thank you for your comment I would like to be treated that way if I disagreed with some now how is social media working we show up you know Remember When Donald Trump used to tweet and then you know you would have one tweet at the top and thirty thousand hate speech below it okay everyone hating everyone some of them hating the president other others hating that person that hated the president others hating everyone right and I'm you know of course the machine is detecting patterns they say okay this first one doesn't like the president let's not show them tweets about the president anymore okay but then the second one doesn't like those who don't like the president let's not show them that anymore right and then eventually the machine comes to the conclusion that humans are rude they don't like to be disagreed with and when they are disagreed with they bash everyone so in the background of the way we're programming them the machines will say okay when they back when they disagree with you with me I'll bash them right this is the typical human behavior now we need to change that we need to change it because interestingly and I use the example of Superman very frequently Superman is that alien being that comes to planet Earth uh with superpowers right those superpowers are neutral they could save our world and they could destroy our world and the difference between being Superman and super villain is the family that raised death raises that being okay so the family Kent decides to tell Superman protect and serve and then we get the story of Superman that we know if you know uh Jonathan Kent I think was his name the father now if he told the child okay you can carry things and break things and see through walls go make me more money go kill everyone that I know is me you know make me richer than everyone make me the master of the world sounds familiar for our you know current life in you know in our hunger for power and capitalism and so on and so forth that's the reality the reality is you would you would use the same superpower and create a very very bad scenario for Humanity I am worried about the machines this is the ultimate suit and I'm worried about the the humans using the machines I'm not worried about them I'm worried about the humans using their machines and and as as those use cases that direct AI to uh to to benefit a few while harming others you know continue to propagate we're going to be in a very bad place now so the good news the good news is the family Kent is not the biological parents of Superman similarly with AI the developer that writes the code is only the biological is only the biological parent but the adopted parent is you and I the ones using the machine so as we use the machine and show how wonderful we are as humans Because by the way all humans are most humans are wonderful inside if we're not hiding behind social media or exaggerated by the mainstream media most of us disapprove of killing most of us want to be loved most of us want to have compassion for our daughters and families and so on right so so there is a lot of good within us if we show that enough one percent of us shows that the machines will actually think that our skin were not scam okay I have some hope I have some hope when you say that that's the uh that's the macro that's the big right and that that gives me some Hope by the way that Superman analogy is outstanding because I someone with um um because I'm not at the 120 range that you discussed earlier so someone at my IQ level can actually understand that um so I appreciate you putting it in that context that's right I wish I were being um let me let's go to some other Solutions let's go to micro I'm listening to this today and I'm like I understand some of this certainly sounds a little bit scary sounds like the world is changing in front of my eyes um or maybe not in front of my eyes that you said something in an interview I was watching where you talked about if you really want to do something hide it in plain sight and that's what's really happening right now but if I'm saying okay I want to protect myself my family my wages my income my quality of life go back to the job thing for a second what should I what's something I could be doing um as an individual listening to this to make sure that my job my career my future is in in my own hands to some extent still what would you say I wouldn't I wouldn't be pursuing these careers I would be doing this these are the skills you might want to be acquiring what would you say to somebody who's I'm sure thinking that listening or watching this it's a very interesting dichotomy if you ask me because while I'm saying AI comes to take our jobs away that's not the immediate term future okay the immediate term future is that someone using AI will take the job of someone who's not right so so you know in the immediate future before AI writes all the code uh you know the developers who use AI better than others to write code will get the jobs that remain so if you lose 10 of the jobs the worst developers the ones that don't have the skill will go then the ones that have the skill but you know AI does it better than them then you know the ones that you know aren't refusing to use Ai and then the ones that continue to use AI will become much more productive and much more capable and so they'll keep their jobs for the near future so what's my immediate answer my immediate answer is jump in and learn those tools okay okay whatever your job is don't resist the wave as a matter of fact ride the wave and while you're re riding the wave do me a favor and deal ethically with the machines right so so show a proper ethical code of being a good human when you're dealing with those machines so that while you're developing and learning and keeping your job or getting a new job you're also teaching the machines to be ethical that's number one number two which I have to admit is a very philosophical but very important conversation we may wake up every morning you and I Ed and everyone listening and think that the world we live in is how it always has been okay it's not at all this is if you just go back a hundred years this is alien in every possible way right and and over you know starting with the Industrial Revolution until today somehow Humanity has identified itself and its purpose would work okay there's nothing inherent within the design of humanity that says without work you don't exist you really think about the original design of humanity the original design of humanity where we connected as a tribe we pondered and learned and developed okay and we simply lived that that was the purpose of life and it's by the way still in a very interesting way most of the spiritual or you know philosophical teachings will tell you that the purpose of life is to live it right we've then in the in the in the capitalist approach to wealth and growth and you know all of the Harvard Business Review uh articles and all of the you know people on Time Magazine and striped suits you know telling you all your purposes to create one more shoe and all of that stuff right we believed that lie and the reality is that this is not us at all our purpose as human as humans if we manage to find our basic needs met okay is to actually live lifefully to explore life fully and that is believe it or not possible with AI so if we manage to get AI to be on our side and I kid you not I'm not making this up we could see a future where you would walk to a tree and pick an apple abundantly you don't have to pay for it and walk to another tree and pick an iPhone okay and both of them because of nanophysics literally cost us the same energy to create okay this is this is not dreaming this is if you understand what we're doing with nanophysics today uh you know it is it's very possible it it's you reorganize the molecules slightly differently it's as simple as that okay now that future is a future where Humanity would go back to the age of nature okay to the age where we actually can interact with life in a way that is human we are not fully human anymore now what does that mean it means that we need to create jobs that depends on the other skill that humans had that is no longer that is not at a threat so remember when we started the conversation we said the two things that created Humanity Humanity as we know it are intelligence and human connection okay intelligence is over it's handed over to the machines they're already more intelligent than most of us and and we're five years away three years away ten years away it doesn't matter from artificial general intelligence but the age of human superiority on intelligence is of okay it's just a question of time so everybody is out there for a second but what he just said was the age of humanity being the superior intelligence is gone just so everybody understands what he said there because there was a little bit of a glitch I want to go back to the work thing just for a second the only place where you and I disagree is that I do think I by the way I agree with you about the greed part but I also do feel that work that in in and I know we've attached value to work but I also like for somebody like myself and for somebody like you my work is my way of expression and I don't want humans to lose their ability to express part of my living is expressing myself part of my work is creating uh the expansion of my being serving other people and so I know what you mean when you said that I just want to make sure that you know that that world you describe I think is beautiful I think people's work is created medications that Keep Us Alive longer that allow us to connect with one another better that so I understand what you mean when you say that um but interesting side of this head is that it you would get the same joy out of it if it wasn't work yeah I think that I think that probably for me I don't view it as work um but I know what you mean when you say it like for me I don't feel like you and I are working right now yeah you are an author and you are expressing something about your book and I am I guess part of one of my careers I'm a podcaster but I don't think either one of us feel like this is laborious and to your point I agree with that about that I think that's going to extend to all jobs okay it's going to extend to all jobs like in all honesty nobody wakes up in the morning who's say an accountant or a you know uh um uh um I don't I don't mean to be against any jobs but there are jobs that are boring like hell right nobody wakes up in the morning and and right and and says my purpose in life is to make the books reconcile right most of us differentiate between you know you and I are the luckiest people to have a job where I can get to meet more amazing human beings and connect and learn and debate and be proven wrong and you know maybe share something that benefits someone it's wonderful but that's not every job okay right and you know you just those messages you get just increased by about 2 million from every account and you just invited we're trying to make AI more loving and friendly and now you've just elicited all these responses from accountants around the world that are going to blast you and now you have to be kind back to them afterwards I I don't I mean I let me rephrase this I wouldn't be excited to wake up wake up in the morning I know what you meant and most people know what you mean but you you follow what I'm saying let me ask you this question so one is that's wonderful advice by the way is to educate yourself and involve yourself in this wave that's here and I have to say I've been remiss in doing that myself and you know I look at guys like me that I'm also a speaker I've watched speeches of me already that are better than me speaking already I've seen this I've listened to music that sounds like Drake that quite frankly sounds better than Drake and so I'm wondering I'm wondering what it's going to do to the world I'm also somewhat I'm very concerned about the ethical part it's interesting that of your diverse background at Google and all the things you've been doing in robotics all your life and all these other things and me having none of those backgrounds we both arrive with your infinite knowledge about this and mine Limited at the same exact conclusion about our concern I am concerned about jobs I'm concerned about the wealth Gap but from a macro from a bigger perspective it's interesting as you step back everybody and you're listening to this and you listen to This brilliant man you know really shine light on something that's right here hiding in plain sight as he says I want to ask you this most it's been on my mind the last few weeks as I asked you to come on the show as I became familiar with you why isn't this the number one story in the world why isn't this on the news I don't care if you're right Media or left media why isn't this in the mainstream this isn't even news on most um social media media any anywhere other than you and a few other people and my only conclusion I can come up with is this technology does allow a again a smaller collection a bigger collection of power for the powerful and perhaps they have an agenda that wants to keep this in the shadows as long as they possibly can so that when it does become a you know pandemic level stuff we go this is beyond our control now we can't do anything about it sorry and there's this um this real small group of people that are even more powerful than they currently are am I crazy am I being am I being a conspiracy theorist when I say that or is there some validity to it I I can comment on the certain part of this okay I I I I I could also nod and say interesting right but let me give you let me give you the solid part of this uh the solid part of this is if you look back at human history uh for the majority of human history since landlords began there has been kings and queens and landlords and peasants right and and the difference between them was automation whatever the automation is so so if you had land the automation process was the land itself the soil you put a seed in by a human you harvest the fruit by a human that's a peasant okay and that most of the wealth goes to the landlord right you have a factory you know the materials come in and a human puts a thread through leather and then you know on the other side a human sells the shoe to someone else and the factory owner and the retail owner and so on is the one that makes all of the wealth okay now the next so call call that say the the soil as an automation we're now starting to create digital soil right and the digital soil is uh you know where you put in a tiny prompt into chair GPT and massive fruits comes out okay and and it's not because you're brilliant you're a peasant it's the machine that is brilliant now there will be landlords and if you really think about it the landlords of AI are the ones that will own the digital soil okay and so there are multiple views of this one view is that it will be the Googles and the matters and you know and the likes the other view is it will be the country that wins because this is an arms race and the third view is it will be the wealthy that created you know if if if there is someone today investing 100 million dollars in an AI that becomes part of that digital soil uh that hundred billion 100 million in the past would return a billion of profits this time it will return 100 billion of profits right so that this is when I talk about the that the differentiation of the uh of the of the Gap but I believe I don't believe in the conspiracies view or of the ability of those people to hide the news okay I think the reason the news is hidden is systemic we have a systemic bias in our system uh you know politicians want to report certain stories uh you know news agencies want to report other stories and this story in itself only lends itself to the system in terms of the system always focusing on the negative and the scary when they talk about the existential risk okay when we talk that you know when when Jeffrey Hinton leaves Google and says I'm warning against the existential risk of humanity that makes news right why because it it sort of warrants more attention because of Humanity's negativity bias than the war in Ukraine okay uh the the the challenge is uh also I don't think any of the reporters any of the politicians any of the actual Business Leaders the investors anyone at all is aware enough to understand the complexity of this story so you don't want to report on things that will make you look like an idiot and and that problem is and I say that with a ton of respect I'm an idiot in a million things and you know but I've lived with those machines I've seen them like my family I stayed in the lab within them I know those machines right and I will tell you this is the story this is it it's not even global warming and climate change this is the story okay this is the most pivotal I I called it in one of my interviews I said this is the Oppenheimer mode this is the nuclear bomb okay and and and the reality is that we I again I try to shy away from the existential risk but this is the first time in history that Humanity created a nuclear bomb that's capable of creating nuclear bombs understand this the machine is now writing machines okay the the the machine that we we think we're prompting it okay but because now so many other software players build agents that are prompting those core artificial intelligences most of the education and data set and training that the machine is receiving today comes from other machines we're now alienated out of that story we've we've Superman landed on Earth and were not even parenting it okay that's where we are and so if you tell if you tell our systemic you know communication methods in the world to communicate that they'll simply say I have no idea what this guy is talking about okay I can't report that story because the system says and and I think you know that about the media the system says there is a pattern to the reporting of the morning show first we're going to talk about a corrupt politician then we're going to talk about the geopolitical issue then we're going to say the economy is going to crush your head then we're going to say a penguin kissed a cat so that at least you can get out of your seat and walk out okay and intelligent people by the way who watch the news if you remove the names and the timestamps it's exactly the same pattern every day it's just you know once it's this politician another it's the other politician you know once it's this economic issue the other is in a different economic issue okay I always I've been watching lately on both sides and I think they're telling everybody what is important and then what to believe about it and then we move on to the next thing and what I'm telling everybody today and you are is this is what's important and we're not really telling you what to believe about it we're telling you to make your own decision but these are the facts engage these are engage in this and and this is the story of our time now let me ask you this last question by the way I've enjoyed today and I this is the one exception on my show where I wish we did go three hours I always respond thank you no I really do because because um because obviously we've scratched the surface here so you've told us that's what the story is today I want you to take your crystal ball out for a second and I don't want to go 10 years forward I want to go five years from now so five years from now what is the story what does the world look like and I and I don't mean what you hope it to be because I heard a lot of Hope in there when we went to the ethics part of how these machines are going to work and and I also saw you wink at me when I asked you if there was a conglomeration of power coming and so I have a I have a uh I have a sense I have a sense that you my sense is that you are um shining the light on what matters now and that there is actually some um you're holding back a little bit to some extent about how deeply concerned you are because you don't want to alarm people but because I want to keep this I want to keep the spotlight on the immediate threats that we have to address fair enough when we address them and I feel stable about them we'll talk about the rest but so let's just let's go five years from now crystal ball it's not that far ahead what does the world look like at that time we would be indeed openly I apologize for using bad language but um unless we start truly putting uh effort in this there will be several uh disruptions that completely redesign the fabric of society as I said jobs is definitely one of them the other which we hadn't didn't have a chance to speak about is AI in the wrong hands so so we are bound to get a significant advantage on one side of the arms race because that's the way AI has been someone finds a breakthrough okay and once you find the Breakthrough look at the you know open AI Google story or alphabet story uh where chat GPT with reinforcement learning gets that immediate advantage that basically puts chair GPT out in the world and for a while the world believes that Google has lost its Edge right and had Google not responded by putting Bard out there you could actually believe that Google would be gone because you know GPT is a very interesting new way of search okay so so you're gonna see that you're gonna see uh some players creating a very big advantage over others and the fear is that this player could be a hacker it could be a defense Authority on one side of the world not not your side it could be a drug dealer that suddenly realizes oh my God there's so much more money if I start to rob banks or convince people or blackmail people or do this or do that and you know and it seems to me that uh Humanity will only create the artificially intelligent policemen on the artificially intelligent criminal shows up okay well that tells me everybody that no no you shouldn't be sorry that's an honest answer and that's why I had you here today because you speak your truth and I want everybody first off Mo I want to thank you number one for by the way taking the risk you're a lot of threat for doing this and yeah yes you are in in every sense of the word everybody and so um not just reputation I'm talking about he's under threat for this and so you know the work you're doing uh May tip the scales in the future of the world and so I'm very grateful for your existence brother and I want to thank you for today and I want to have you back because this is worthy of more than just the time today but we did accomplish what I hoped we would today which was shining a light on all of this and by the way everybody that's why you want to go get scary smart the future of artificial intelligence and how you can save our world go get that book by the way and after you read it and you feel like we're in deep then go read Saul for happy it's the perfect handle engineer okay with what's happening it'll be okay and that's his other book and um and I just really really want to thank you today and everyone you have an obligation now for your family for wherever you live in the world to begin to educate yourself to give you involve yourself with this technology read about it stay close to the sources that provide you any information about it keep yourself educated and as on The Cutting Edge as you possibly can and the people that are in power that are around you let's start to get them to do you know some discussion about this and shine some light on this topic because the world is changing right now in this notion that um the machines are already more intelligent than humans is of great concern so all right everybody Mo thank you so much for today I enjoyed this tremendous having me it's it's really kind of you to to put me on your platform and I'm holding you to showing me around Dubai when I come out there on October absolutely on me coffee is on me and you will love it I love it brother all right God bless you everybody God bless everybody else here max out your life share this episode if you've ever gonna share one it's this one share it everybody take care [Music] thank you

Channel: Ed Mylett

Views: 79,804

Keywords: ed mylett, mylett, ai, artificial intelligence, oppenheimer moment, oppenheimer ai, google x gawdat, mo gawdat, ai dangerous, benefits of ai, maxout, powerofonemore, the power of one more

Id: Ae4o92F4tVM

Length: 70min 0sec (4200 seconds)

Published: Tue Aug 08 2023