WEBVTT 00:00:01.120 --> 00:00:04.797 Google's rise to becoming the most powerful company in the world is 00:00:04.880 --> 00:00:07.917 gripping and all-encompassing. 00:00:08.000 --> 00:00:12.157 >> The assistance around us, the knowledge that we have at our fingertips, 00:00:12.240 --> 00:00:18.840 literally will be profound. It'll be a very, very powerful world. 00:00:22.240 --> 00:00:24.557 For a decade, Eric Schmidt operated as 00:00:24.640 --> 00:00:27.117 its CEO and afterwards as the executive 00:00:27.200 --> 00:00:29.917 chairman. Now he works on a series of 00:00:30.000 --> 00:00:34.157 projects and investments as well as advising governments on applications of 00:00:34.240 --> 00:00:37.517 AI. Mr. Schmid, welcome and thank you very much for the interview. 00:00:37.600 --> 00:00:40.957 >> Thank you for having me. >> You advise governments on the 00:00:41.040 --> 00:00:44.717 implementation of the AI. What are they most interested in? 00:00:44.800 --> 00:00:48.077 >> Well, governments are concerned about two things. One, they want to win, but 00:00:48.160 --> 00:00:52.797 they're also concerned about safety for their populations and can it be misused? 00:00:52.880 --> 00:00:57.357 You have to solve both problems. They have to both invest but they also have 00:00:57.440 --> 00:01:01.517 to regulate and they're trying to sort it out. Some are better than others. 00:01:01.600 --> 00:01:05.517 >> There is huge excitement about AI at this conference. However, when you 00:01:05.600 --> 00:01:09.517 leave, you have the feeling that people are actually really scared. Scared about 00:01:09.600 --> 00:01:13.517 their jobs, scared about AI going into the wrong hands. How valid are those 00:01:13.600 --> 00:01:14.477 fears? 00:01:14.560 --> 00:01:17.757 >> Uh there's some truth to them. Um the 00:01:17.840 --> 00:01:21.117 models as they get smarter, if you will, 00:01:21.200 --> 00:01:25.357 have dangers embedded in them. The industry puts what are called guardrails 00:01:25.440 --> 00:01:29.517 on them to try to prevent them from answering questions like how do I kill 00:01:29.600 --> 00:01:31.197 myself? How do I kill somebody else? 00:01:31.280 --> 00:01:35.437 Things which are obviously dangerous. So governments have trust and safety 00:01:35.520 --> 00:01:38.157 programs to do that. Remember that this 00:01:38.240 --> 00:01:40.717 is a wave that is a generational wave. 00:01:40.800 --> 00:01:43.197 It will go through it will take 10 or 15 00:01:43.280 --> 00:01:45.917 years. The PC revolution was like that. 00:01:46.000 --> 00:01:47.997 The internet revolution was like that. 00:01:48.080 --> 00:01:52.237 Social media was like that. AI will be the same. We're just at the beginning of 00:01:52.320 --> 00:01:54.637 the changes that AI will bring in 00:01:54.720 --> 00:01:56.957 medicine, in science, in conflict, in 00:01:57.040 --> 00:01:59.357 information, in misinformation, in what 00:01:59.440 --> 00:02:04.237 you do and what I do. It will all change mostly for the better. 00:02:04.320 --> 00:02:07.837 >> Where are you most concerned about when it comes to AI and SAI? 00:02:07.920 --> 00:02:09.597 >> I'm actually concerned about us missing 00:02:09.680 --> 00:02:12.557 the opportunity. And the reason is that 00:02:12.640 --> 00:02:15.517 the arrival of an intelligence that's 00:02:15.600 --> 00:02:19.757 sort of like a polymath, one of these sort of brilliant minds that is aligned 00:02:19.840 --> 00:02:23.837 with you, that works with you, is a major change in human productivity and 00:02:23.920 --> 00:02:26.317 efficiency. And we have to harness that 00:02:26.400 --> 00:02:29.197 power. And we harness it for education, 00:02:29.280 --> 00:02:31.997 right? for medicine, for the efficiency 00:02:32.080 --> 00:02:34.397 of economies, for generating wealth, for 00:02:34.480 --> 00:02:39.197 making the middle class stronger, from taking people from absolute poverty into 00:02:39.280 --> 00:02:41.277 the middle class, for climate change. 00:02:41.360 --> 00:02:45.517 All of that is bene benefits from AI if it's done correctly. 00:02:45.600 --> 00:02:49.357 >> You were at the helm of Google when social media came along. Did you see 00:02:49.440 --> 00:02:53.517 back then the negative implications that we are so aware of now? 00:02:53.600 --> 00:02:58.077 >> I did not and I can confidently say that almost all of my peers did not see it 00:02:58.160 --> 00:03:02.397 either. Um when I think about it, I think about it in terms of the bubble 00:03:02.480 --> 00:03:04.637 that we live in. All of us are similar. 00:03:04.720 --> 00:03:08.157 We have a certain level of education, a certain set of political views. We come 00:03:08.240 --> 00:03:12.557 from the west. And I think we did not understand how these incredibly 00:03:12.640 --> 00:03:16.957 important tools which benefited many many people could also be misused. I 00:03:17.040 --> 00:03:18.957 think that the fact that social media 00:03:19.040 --> 00:03:22.237 has enabled tribalism and misinformation 00:03:22.320 --> 00:03:26.877 was not in the plan. That's not what we designed, but it happened and now we 00:03:26.960 --> 00:03:30.637 have to address it. How can we learn from this example when it comes now to 00:03:30.720 --> 00:03:33.437 the AI revolution? >> Well, I've written one book and I have 00:03:33.520 --> 00:03:37.757 another book coming out in in November on the same subject. We have to get 00:03:37.840 --> 00:03:39.997 ahead of AI and its impact on human 00:03:40.080 --> 00:03:43.597 life. When your child or your grandchild 00:03:43.680 --> 00:03:47.117 has as its best friend a nonhuman actor, 00:03:47.200 --> 00:03:51.437 right, a voice on a computer and they're five years old, what does that do for 00:03:51.520 --> 00:03:55.357 development? Should we allow it? How should it behave? Children are very 00:03:55.440 --> 00:03:59.997 important. um all sorts of issues around messages and identification. I 00:04:00.080 --> 00:04:04.557 personally am in free in favor of free speech for humans but not for computers. 00:04:04.640 --> 00:04:07.997 I want to hear from humans and their voices. I don't want to hear what 00:04:08.080 --> 00:04:11.837 computers think humans think. I want to hear from actual humans. Those are 00:04:11.920 --> 00:04:15.437 simple principles. Would not be difficult to write them down. 00:04:15.520 --> 00:04:17.677 >> You have spoken about implementation of 00:04:17.760 --> 00:04:20.797 AI in warfare just now. How much of AI 00:04:20.880 --> 00:04:23.597 is currently used on the battlefield? 00:04:23.680 --> 00:04:26.557 almost none is used on the battlefield 00:04:26.640 --> 00:04:28.477 and most of that I think is good. 00:04:28.560 --> 00:04:30.637 Battles are very rough. It's very tough. 00:04:30.720 --> 00:04:34.077 AI has every reason to in my view to 00:04:34.160 --> 00:04:36.397 become an even more lethal solution in 00:04:36.480 --> 00:04:38.477 warfare. And the wars that I have seen 00:04:38.560 --> 00:04:42.477 are horrific. So unfortunately war will 00:04:42.560 --> 00:04:46.637 be changed by AI too. Um in some sense it's good because it'll become more 00:04:46.720 --> 00:04:49.117 targeted but it's bad because it'll 00:04:49.200 --> 00:04:53.757 become easier. But do you feel that investments in warfare a government 00:04:53.840 --> 00:04:57.357 should also change and be more adapted to the new world? 00:04:57.440 --> 00:04:59.997 >> Well, I do because the simple simple 00:05:00.080 --> 00:05:04.797 answer is that drones which are inexpensive can take out a tank, right? 00:05:04.880 --> 00:05:06.557 So tanks are very expensive. Drones are 00:05:06.640 --> 00:05:11.837 very cheap. Tank versus drone, the drone wins. So all of these assumptions that 00:05:11.920 --> 00:05:15.357 militaries have made about the importance of large things like aircraft 00:05:15.440 --> 00:05:18.477 carriers and submarines and army bases 00:05:18.560 --> 00:05:22.557 and so forth and tanks and other very exotic weapons uh which you see in 00:05:22.640 --> 00:05:26.637 America all the time are are essentially being undermined. All those assumptions 00:05:26.720 --> 00:05:30.557 are being undermined by this new technology as it rolls out. I'm 00:05:30.640 --> 00:05:35.117 confident that in a decade the the the structures of militaries will be very 00:05:35.200 --> 00:05:38.877 different for that reason. Here in Saudi, we have a sense that there is a 00:05:38.960 --> 00:05:42.477 lot of pushing towards leadership of this AI revolution. How well is the 00:05:42.560 --> 00:05:47.437 kingdom placed to do so? >> Saudi is very well positioned if it does 00:05:47.520 --> 00:05:51.597 a whole bunch of things. Saudi has abundant energy, obviously, abundant 00:05:51.680 --> 00:05:55.837 land, and a strong commitment to building things. U the world is going to 00:05:55.920 --> 00:06:00.077 need gigawatts and gigawatts of data centers. These are very, very large data 00:06:00.160 --> 00:06:04.877 centers. They could be in Saudi and Saudi has good enough connectivity that 00:06:04.960 --> 00:06:08.957 Saudi could be one of the places not the only place but one of the places where 00:06:09.040 --> 00:06:13.437 AI technology is invented and developed often by people who are in other 00:06:13.520 --> 00:06:17.197 countries using them. The key to this remember is that the internet is 00:06:17.280 --> 00:06:21.037 borderless in that sense and the work can be done here while the people live 00:06:21.120 --> 00:06:24.397 in the countries where they are are from. It's a good outcome. 00:06:24.480 --> 00:06:28.237 >> You did say that a few things would have to be done like what? Well, the most 00:06:28.320 --> 00:06:32.237 important thing is a commitment to do it. And the good news about this 00:06:32.320 --> 00:06:36.157 government is this government is willing to put large amounts of money behind 00:06:36.240 --> 00:06:40.477 successful investments. One thing I have said to the Saudis this trip is those 00:06:40.560 --> 00:06:44.877 investments will be very, very good financially. It's good business. It's 00:06:44.960 --> 00:06:49.277 good strategically. And frankly, it's good for a Saudi as it's trans goes 00:06:49.360 --> 00:06:53.037 through its transformation to become a world leader to be at this level. They 00:06:53.120 --> 00:06:56.157 need to do this to get to where they want to be. 00:06:56.240 --> 00:07:00.077 >> Here at FI8, as we said, there is so much excitement about the AI. 00:07:00.160 --> 00:07:02.877 Realistically, where are we as a society 00:07:02.960 --> 00:07:04.957 with implementation of AI? 00:07:05.040 --> 00:07:07.917 >> We're just at the beginning. Um, you as 00:07:08.000 --> 00:07:12.717 a consumer see it in the way the social media apps work and Google works and so 00:07:12.800 --> 00:07:16.797 forth. But I want you to think differently. Today in your show, you 00:07:16.880 --> 00:07:20.557 have a producer and a director and a whole team that helps you. Why is that 00:07:20.640 --> 00:07:24.957 not all those functions not handled by your AI assistant that knows exactly 00:07:25.040 --> 00:07:29.757 what's going on? Helps you become even better than you already are. Each of us 00:07:29.840 --> 00:07:33.197 will become better at what we do because we'll have these assistants and these 00:07:33.280 --> 00:07:37.437 assistants will be with our permission highly personalized to what we do and 00:07:37.520 --> 00:07:38.877 what we care about. 00:07:38.960 --> 00:07:41.517 >> Do you feel that educational system is 00:07:41.600 --> 00:07:44.797 prepared to teach an entire new mindset 00:07:44.880 --> 00:07:47.677 of this future work? What I would say is 00:07:47.760 --> 00:07:50.237 that AI transforms every industry 00:07:50.320 --> 00:07:54.717 including education. Um it's obvious that teachers will be important. 00:07:54.800 --> 00:07:56.957 Teachers with AI will be better teachers 00:07:57.040 --> 00:07:58.397 than teachers without. 00:07:58.480 --> 00:08:01.917 >> And last question, how important with 00:08:02.000 --> 00:08:06.157 all the AI systems we have, is it still to sit down and talk like we are doing 00:08:06.240 --> 00:08:10.637 here today, like all these people are doing here today at FI? 00:08:10.720 --> 00:08:14.317 >> I think that human contact, human connection is going to be with us for 00:08:14.400 --> 00:08:18.797 the rest of humanity. Um eventually the computers will run a whole bunch of 00:08:18.880 --> 00:08:23.197 stuff while we're still talking to each other. Um so I think human behavior is 00:08:23.280 --> 00:08:25.357 not going to change very much but the 00:08:25.440 --> 00:08:27.917 assistance around us the knowledge that 00:08:28.000 --> 00:08:30.637 we have at our fingertips literally um 00:08:30.720 --> 00:08:35.517 will be profound. Think about what you know now from your use of Google right 00:08:35.600 --> 00:08:38.877 compared to before Google and now imagine when you have an assistant 00:08:38.960 --> 00:08:41.197 that's using Google or from Google that 00:08:41.280 --> 00:08:43.757 can really uh upcharge what you do. 00:08:43.840 --> 00:08:46.157 It'll be a very very powerful world. 00:08:46.240 --> 00:08:47.277 >> Thank you very much indeed. 00:08:47.360 --> 00:08:50.440 >> Thank you.