FOR EDUCATIONAL AND KNOWLEDGE SHARING PURPOSES ONLY. NOT-FOR-PROFIT. SEE COPYRIGHT DISCLAIMER.

Former Google CEO Eric Schmidt Talks Future of AI

current technology seems to have infinite possibility that so many of us can’t even begin to understand my guest today is someone who’s seen it all evolve he’s an innovator who has shaped the technology that we use as former Google CEO and chairman he pioneered the company’s transformation from a Silicon Valley startup to a global leader his many board positions include chairman of the US Department of defenses Innovation board chairman of the US National Security Commission for artificial intelligence he has written many books and he’s the global leader in AI was recently kned and is’s here to talk to me today about ai’s infate possibilities as well as the hugely fastpac the technology companies are moving Eric Smith thank you so much for joining me it’s such a privilege to have the opportunity to speak to you tally when you invited me I immediately said yes because of who you are and what you’ve done in Australia and everywhere so I’m really happy to be here you are very very kind and now as someone who has a better understanding of just how quickly technology is changing and what the capacity and where it can go is you’ve mentioned that your stck by just how fast the pace of tech companies are moving can you tell me about that so one way to say it is if you watch media you’re on social networks everything seems to be all screwed up in the west and maybe the world everyone’s unhappy about this or that and it doesn’t seem like a lot of progress on the important issues is being achieved that is not however true in Tech I am really struck by how fast this next set of changes I think of it as everything everywhere all at once and the the ability and the tools that are being invented here will shape our our not only Society but our medicine the future of History war and so forth in ways that I think are are largely unimaginable and I know it’s going to sound crazy but I would tell you that this change is underhyped not overhyped the real question is whether societies globally are ready for it which is something we can explore with with these changes you’ve spoken about how they can cure cancer how they can help climate change how they can make energy use better what are you most excited about with all these incredible things well in my World which is largely the tech World there is what I call the San Francisco school and and pardon me for a minute as I Define it these are people who believe that one or two or three more of the crank of what they’re doing which I’ll explain in a sec the training the models will produce intelligence that’s equal or better than the sum of human intelligence they actually believe this I personally think they’re a bit optimistic but directionally I think they’re correct and so the arrival of a nonhuman intelligence that’s helpful to to us all of us is a very big deal so for you as a presenter an author the writing and the the work you do imagine if you have the perfect assistant who writes everything suggests what you do ints what you’re good at figures out what your audience really cares about and serves both as coach Mentor assistant and essentially your partner and that’s priv it to you literally on your phone and now imagine that everyone in the world has such an assistant in my case one of the things that’s happening is that I’ve been a software programmer and a computer scientist for my whole life 50 years or more and during that time programming was everything these computers look like they can replace me can they is is this something we should be worried about like are we gonna all be replaced well I I think humans will continue but my job as a sort of okay programmer is probably better done by a computer with me supervising it it’s going to make fewer mistakes it’s going to get things better it’s going to work better with other program or computer things and one of the things about the San Francisco school is people believe that at the moment of course it’s at this enormous pace of Discovery and they actually believe and I’m kidding you not that the technology will develop its own AI scientists and the AI scientist will be faster than the human scientist at Discovery in new things so an example would be a system that makes proofs and another one that checks them well it can just generate new knowledge it can generate new physics it can generate new understanding of chemistry and and going back to things that bevil all of us Health Care drug drug Discovery we there’s a company that’s using Quantum simulation to take existing drugs and make them longer more effective and more and and have fewer side effects that works today even without quantum computers so the reason to think that this transformation is just at the beginning is to look at where all the money is and there’s enormous amounts of money thousands and thousands hundreds of thousands of people working on this stuff um in fact one of the interesting concerns that one of my friends said last week is they they said that their modeling of the power needs are that America will run out of of power literally energy by 2028 four years at the current rate is way to fix that of course well build more power plants but I have a better idea I tell people think Australia now what’s true of Australia aside from being far away we really like Australians they’re really great people and they’re also part of our national security they’re part of the five eyes so maybe instead of building the data centers in America the Australians can organized to become a global Training Center given the incredible amount of solar W solar power wind and so forth that and natural resources that your country has imagine if instead of training here in the United States where we’re running out of power all the stuff is done globally but done in machines that are in Australia this is a reasonable scenario where five years ago I would have said it’s not possible at all well I like the sound of that and it means that we get you to Australia it sounds so incredible what are some of the challenges that you can foresee I mean apart from running out of electricity but when when it comes to problems with the AI are there any that you foresee well I think I can scare people to death with a couple of scenarios and I’m not trying to but let me let me talk about the ones that we know about we want to make sure that these systems do not have access to weapons right imagine the computer decides that part of its programming is to simplify its World by getting rid of a few people and then it decides oh why don’t I get rid of all of them so all of those scenarios need are obvious and need to be protect prevented another one is biology We Now understand with synthetic biology how to build viruses that are really quite harmful to many people or everyone and so policing that is important the scale of cyber attacks that are possible give you another example in this San Francisco scenario I was describing that computer may be be smart enough we don’t know yet to break in and break through all of the controls that prevent you or me from breaking into the financial system so what would it do well maybe it would just steal the money for itself because it decides computers need more more money which is of course Preposterous so there’re all of these questions that we don’t know another example is misinformation um it’s pretty clear to me that the the Govern that misinformation is going to really screw up Democratic politics because people are trained to believe what they see and they hear and yet these computers will be capable of building just extraordinarily good um video and audio that’s fake that’s a problem um cyber attacks I mean we can just go on um to another example you have a child and your child’s best friend is um a bear you know a little bear toy and every year you give the child a new a new bear which is smarter than the last version bear and the child gets smarter too so they become best friends we don’t have a model where the com where where your children’s best friend is not a human right we have to have these conversations indeed Dr Kissinger and I wrote a book about this and we have a second book coming out in November um he unfortunately died at a half but he spent his last week before he died finishing the book that’s being published on the anniversary of his death called Genesis it’s called Genesis and it’s fundamentally about the arrival of a new nonhuman alien intelligence and how do we interact with it how do we make sure that it does what we want now my standard line of this is pretty simple when the computer decides to do something on its own do you know what I think we should do I think we should unplug them literally unplug it and the reason is if it starts to talk in a language to other computers that we don’t understand and we don’t know what it’s doing unplug it can we always will we always be able to do that until the computer figures out to evade that by coming up with other power sources but at least for a while but if you remember two rules which is first think Australia right great Partners lots of power lots of smart people and unplug it as a rule if it gets out of hand then that’s a kind of simple model of the next 5 years okay that’s good rules thank you now whilst we’re sort of touching on the potential Doom and Gloom there are multiple Wars currently raging in the world how will AI change the way that Wars are fought we you’re seeing this in Ukraine I’ve been helping Ukraine for a long time and because of the ex exigencies of war and in particular the dire situation that the country F finds itself in it’s become an innovation nature there are more than 600 drone companies and what they’re doing because they don’t have an Air Force and they don’t have a Navy right they for forgot that in the separation of the Russians I guess um but they have to counter the Russian attacks and the Russian missiles they’ve invented incredibly new interesting Battlefield techniques um typical example would be long range drones that destroy stationary and moving targets and they’re they’re doing things across the front line that were thought to be imp possible now the Russians by the way um who I am obviously not happy with are also very clever and they’ve invented electronic Wars walls to block that so if I look at Australia or the United States or the other very tight Partners In The West we’re not prepared for this we our weapons don’t necessarily work against the new Russian T tactics and you can imagine I’ll simp give you a simple example a $5,000 drone car in 2 kilograms of of U ammunition can destroy a $5 million tank now the US has thousands of Abrams Tanks lying around not doing anything what are we going to do with them they’re useless right I’d much rather have 5,000 drones but if I have 5,000 so I could have five million sorry 5,000 five million drones for the same price as the Thousand tanks so you see that the cost benefit of drones and automation think of it as robotic War which is scary so I’m not endorsing it I’m just saying that’s what’s happening and we’re going to face issues there too so uh I led this uh Commission on AI and National Security in the United States which was well-received and we distinguished between automation under control by humans which we’re in favor of so you have a soldier or a leader or a President Who authorizes something and it happens automatically versus automatic weapon systems which just respond and make their own decisions and that’s terrible there’s a movie called Doctor Strange Love from 1964 where you see the consequences of of course that’s a farce I don’t want that to be real and is this I mean you have advised the government is is this something you advise them on is it something when you know when you talk about Russia and Ukraine and and how America isn’t at that level is that something you have advis and something that America and Australia can get up to date on to be able to fight a war in that way and if necessary um they can they’ve chosen not to because they’re still in the old Paradigm and and I’m sorry to be blunt here but these systems in the west which includes Australia but it’s certainly not probably copied from the US or organized around buying incredibly expensive very precise very very vulnerable systems example would be aircraft carriers why do you need an aircraft carrier it’s a big sitting Target against Lots and lots of missiles they’re hard to defend I’d much rather have for the 18 billion that an aircraft carrier is 18,000 that’s a million dollars let’s say aund let’s let’s in fact say 1.8 million drones at $10,000 a piece I could do a lot more damage in a war in my view with automation so the whole mindset from Australia and the us and our other partners has to change give you another example let’s think about the Israeli Hamas War and what Hamas said is terrible it’s horrible it’s a war crime wouldn’t it be great if the you if the Israelis had had drones which would immediately hit their their missiles and would also do tunnel clearing automatically that is without endangering the IDF soldiers so of course this technology works you just have to make it work you have to make it work at scale you have to deal with the various security issues but my point is that the arrival of automatic war in the sense of of autonomy and coordination and having the person I’ll give another example I was sitting in Kei one day and they were running a mission and they have the people the soldiers are in uniform they’re sitting in front of a PC they’re drinking coffee and they’re running the war right at and and they’re safe in Kei that’s the future of this technology and uh I strongly believe that the Australian government should lead in this area because when these Technologies are invented they’ll be highly exportable as a business people are going to want the technologies that that can protect them there’s so many examples where uh both from threatening of drones as well as think of it as robotic war in the air and on the sea how do you defend it how do you proper it my own view which is just my and maybe seen as naive and if so I apologize is that eventually both sides will have this and it’ll be very hard to have a land War now Australia doesn’t have land Wars because you’re surrounded by an ocean but most countries have a border and often they have a border of somebody they don’t like it’d be really nice if you if one side couldn’t invade the other by land in my view that will happen I I hope you’re right and I hope the powers that be listen to your advice um when it comes to the incredible capacity that AI has how important and how possible is it that everyday people can learn proper commands to be able to use it effectively well I I’m going to make a different argument um in my life I spent most of my time Building Products that no normal human could use and because I was such a specialized person it didn’t occur to me that my stuff was not usable by any normal person who’s trying to have a normal life and get their job done and I think our our industry went through these sort of weird specialized things I mean I used to have to to reprogram my modems and things like that that you would never do today we’ve learned now with UI techn you user interface Technologies to anticipate to do things automatically so an example would be my favorite example is you see all of these people doing influencer things where they’re doing videos of themselves doing funny things to post them why doesn’t the computer generate that why do I have to fill myself thank God I’m not doing that I’m too old um but but why doesn’t the computer do that on your behalf right so why doesn’t the computer anticipate what you need and manage it so I’m arguing that in fact the the industry which has grown very quickly is on the cusp of automation of a good kind where it sort of understands what you need um and and it just sort of works so an example for me is I move around a lot and I have to switch networks and so forth and cellular networks and Wi-Fi why do I have to know that why doesn’t the computer do all that automatically and when you start making a list of things which are annoying and are waste of your time as a user uh I strongly encourage everybody to demand your vendors fix it like why do I have to manually install the app why can’t you manage it assuming that you have good security and you know who I am which we can do quite well now why doesn’t it’s just sort of self-organized in the way that Eric or tally or whoever likes I love it it now one of the issues that you mentioned in your book is the Divide between societies who can afford to adopt new technology and those who can’t how do you see this manifesting and changing the world well let me state it as a worry if I look going forward the technologies that I’m describing will be invented in the US and in some of our partners which will include Australia and in China and some of her Partners it’s unlikely they’ll invented in Russia North Korea you know the bad places because of sanctions it may it there’s a good chance it’ll be invested invented in India as well I’ve just named five or six countries but there’s 197 countries what happens to the rest of them are they forced to become surfs if you will to these Giants um what happens to their culture what happens to their citizens what happens to their Educational Systems what happens to their language and so uh I worry about countries that are sort of middle- inome countries that are not that well run their universities aren’t that strong we know the people are incredibly smart we know the human Talent is there it’s being pushed down by bad political decisions or cultural decisions or religious problems or what or war or whatever I’m not an expert on that so I worry that those countries will be left behind because they’re not big enough or they don’t partner right or they don’t have strong enough technology to participate in this revolution so you see this in the political world where globalization has produced winners and losers and the losers don’t like it and the losers are fighting hard why can’t we come up with a way where these Technologies benefit everyone now if you look at the last boom it was true that the rich got much richer but the poor also got better we have to figure out a way to make sure that these Technologies generate economic and cultural benefits for everyone and what does the average person care about cost of living education for their children safety um the future of everything their health their health care we we we’ve not fully in my view embraced all of that and that’s a problem to be solved to some degree this occurs because the tech companies of which I’m obviously a member are optimized around uh shareholder value economic growth wealth and so forth and the constituencies for that everybody else are not as strong as those constituencies that’s a problem to be fixed do you think it will be able to be fixed do you for say that um there’s certainly a way and I’ll give you an example um I think we should set a goal globally of having an AI doctor for the world and an AI tutor for the world world so the AI doctor is not a doctor are you gonna be the doctor not a chance an AI doctor is really the computer that helps the nurse practitioner the healthc care provider especially in the developing world where you typically have a nurse as your Healthcare proor those people need a lot of help you could really especially if it were in the local language now let’s think about education why do we not have a tool which is a piece of software on the students phone which essentially engages them in their local language is one of infinite possibility we know that IQ is uniformly distributed around the world we know that in every one of these wacky countries that I refuse to go to you have true Geniuses who are stuck there can we give them a path out and that starts with education and health and this can be done relatively inexpensively if we worked on it I don’t understand how we could be in such a situation in my industry where we have so much wealth but we haven’t done that hardly now you’ve spoken about challenges for small companies because of the sheer cost of developing the type of software what do you look for when you invest in startups well in my case I try to invest in raw startups and raw startups are wacky because you fundamentally have no idea what they’re going to do so you’re invest in the idea but more importantly the person and their ability to build a team if they can if they’re very sharp they understand their their target well and they’re willing to work really hard they can create great wealth that’s the lesson of this last generation of of Technologies and I’m quite serious about is there something that particularly excites you at the moment and something you’ve invested I know you have a lot of Investments um well most people are focused on the enormous wealth creation that’s occurring from the big tech companies one of my friends a couple of years who’s an actual professional investor said he had calculated that the vast majority of the wealth would actually go to the existing companies and not the new startups and that turns out to be true and in my case I’m very interested in deep Tech so for example Fusion Energy I spent some time today working with a fusion company and fusion would solve a lot of the energy problems in the world and it would certainly help solve the climate change problems and again the new techniques may allow that there are a number of interesting startups there amazing I’m also very interested in video I think that video um and of course this is your industry so you under extremely well I’ve always assumed that I would just be typing forever but most people actually like video and they like to communicate with video and video messaging we are now very comfortable zooming and FaceTiming and all that kind of stuff it looks to me like the the new answer to every question is video and shortform video medium- range video long-term video and video that’s also generated as well as a real video so you can imagine a world where a lot of your information space an example would be today if I use Google Google will give me a Gemini generated summary so you can imagine in text you can imagine that in the fure future it’ll give me a AI generated video which is even more entertaining to communicate the things there’s every if you look at Sora from open AI you look at these other stable diffusion and so forth they can do such photorealistic images um that it’s it’s a a visual Delight right um it’s sort of like looking at the Sydney Bridge um and the harbor and the Beautiful lights in the Opera House you look at it and you go that is so incredibly beautiful and I think we can do that literally on steroids picture after picture after picture and video and I think people will be measure mesmorized by that technology and it looks like it’s coming it’s too expensive right now to do it at scale but if you look at the cost um improvements in performance they’re very significant and U one of the remember that the iPhone 1 barely worked and today the iPhone 16 is the supercomputer in your pocket Within enormous power um I was recently looking at the where I went to college my phone is a 100 million times more powerful than the computer that the whole university used when I was in college wow we forget how fast this stuff is happen that is incredible um Elon Musk recently got into a war of words with our Australian prime minister over regulations with X if a company does get banned can people actually use llms to get around Banning and regulations um I’m not familiar with that specific dispute so let me not comment on any of of the Elon part or the Prime Minister part um who I’ve heard is quite good um but I’ve not met them so um llms are not currently current so they’re trained typical scenario they’re trained for six months and then they do fine-tuning as it’s called for six months and then they’re released so part of the current problem is that the llms are not you know today’s news they’re they’re very very powerful they’re very very insightful but an example is I asked who’s going to win the Oscars this year and it gave me the answer from two years ago um so that’s that’s the limitation of the technology uh the consensus in the industry is this is going to get fixed in the next one to two years and the way they do it by the way is continual training when I started at Google Google had a fixed one- Monon crawl and then the engineers figured out a way to do continuous crawls and so all of a sudden Google became current so I would expect the same thing with LM okay thank you now as someone who has not only shaped so many things that are integral to our modern society but also inspired so many people who inspires you well on a technical basis um Elon is a truly brilliant entrepreneur and when I say that I mean the ability to pull off the amount of money and the scale and the complexity of the systems that he’s building using SpaceX is really a remarkable achievement um from my perspective my heroes for a long time were Larry and Sergey who were the founders of Google who I work with every day and so I think my my today answer is probably if you look at the tech leadership whether you like them or not is not the point if you look at Mark and Elon uh Larry and Sergey you look at SATA at Microsoft you look at Andy jasse at Amazon they are operating systems at scale that we would die for in any other area imagine if the prisons were run as well as these companies are if the government services were run as well as these I’m not arguing for that I’m just making the analogy so one of the observations which is uncomfortable I think for most people is that the corporate Corporation structure with a CEO and a board and the incentive structure produces enormous progress and much more refined I’ll give you another example using Elon starlink those satellites are moving get the map look at the speed that they’re working on and yet the handoffs are seamless if you’d asked me a year ago I would have said it’s impossible in that particular case they bought a company that had been working on it for 10 years and they capitalized it and they got it to work Amazon as you know is entering the market with somebody called Cooper um and I’m assuming the Chinese will do something similar um uh so that’s a good example where this kind of innovation not only do is it extraordinary to look at starting with like the Tesla is an example but it inspires a whole industry through competition to catch up we all benefit from that Innovation and the subsequent competition among the key players amazing Eric thank you so so much for joining me I can’t wait to read your new book and I hope that we get to see you in Australia very very soon um and I promise to come and tell you’re amazing I’m really happy to do this and as I said every time I get a chance in the US government I say think Australia great Partners lots of power great place to do a lot of AI training so thank you guys very much I’m personally going to hold you to that promise to come to Australia thanks

FOR EDUCATIONAL AND KNOWLEDGE SHARING PURPOSES ONLY. NOT-FOR-PROFIT. SEE COPYRIGHT DISCLAIMER.