WEBVTT - The AI Arms Race Part Two 0:00:00.200 --> 0:00:07.280 Brought to you by Toyota. Let's go places. Welcome to 0:00:07.440 --> 0:00:15.480 Forward Thinking, Hey darn, Welcome to Alward Thinking about podcast 0:00:15.600 --> 0:00:18.639 that listen the future and says Einstein can't be classed 0:00:18.640 --> 0:00:22.080 as witless. He claimed Adams were the litt list. I'm 0:00:22.160 --> 0:00:25.400 Jonathan Strickland and I'm Joe McCormick, and our other host, 0:00:25.520 --> 0:00:28.040 Lauren Vocal bamb is not with us today. She is 0:00:28.120 --> 0:00:31.920 in New York doing very exciting things. But today we're 0:00:31.920 --> 0:00:36.440 gonna be initiating part two of a two part episode 0:00:36.920 --> 0:00:40.760 on the AI arms race, and the last episode. If 0:00:40.760 --> 0:00:42.360 you haven't heard that yet, you should go back and 0:00:42.400 --> 0:00:44.199 listen to that one first, where we lay all the 0:00:44.240 --> 0:00:46.520 groundwork for the stuff we're gonna be talking about today. 0:00:46.920 --> 0:00:49.320 Last time, what do we do? We talked about sort 0:00:49.320 --> 0:00:53.760 of some definitions of different concepts in artificial intelligence, how 0:00:53.840 --> 0:00:57.400 likely we are we think to achieve them, what ways 0:00:57.480 --> 0:01:01.880 they might be achieved, some potential stumbling blocks to achieving them. 0:01:01.920 --> 0:01:05.320 But in the end we want to say, today, okay, 0:01:05.400 --> 0:01:09.200 let's assume that people create what's known as an artificial 0:01:09.360 --> 0:01:14.399 general intelligence. Is that going to lead to a worldwide 0:01:14.640 --> 0:01:20.520 arms race for intelligent machines? Boy, that's spoiled it. No, 0:01:20.680 --> 0:01:23.479 We've got a lot to say about this, and this 0:01:23.560 --> 0:01:25.640 is if you listen to the last episode you heard. 0:01:25.640 --> 0:01:28.400 At the end, I mentioned that we have some interesting 0:01:28.480 --> 0:01:33.039 characters to talk about, people who have perspectives on this idea, 0:01:33.160 --> 0:01:36.440 and the first person we're going to chat about falls 0:01:36.440 --> 0:01:40.280 into that interesting character category quite handily. Yeah, so this 0:01:40.360 --> 0:01:42.479 is just one example. There have been lots of people 0:01:42.520 --> 0:01:44.520 actually who have written about the idea of a looming 0:01:44.560 --> 0:01:48.640 AI arms race. But one example of a guy with 0:01:48.920 --> 0:01:54.080 this prediction is a guy named Sultan Istvan who is 0:01:54.240 --> 0:01:59.040 the presidential candidate in the United States for the Transhumanist 0:01:59.120 --> 0:02:03.320 Party ticket. He's a trans humanist journalist and I guess, now, 0:02:03.400 --> 0:02:08.320 politician I suppose, yes, enthusiasts seems too modest a word. 0:02:08.840 --> 0:02:11.880 He actually he had on his website a a suggestion 0:02:11.919 --> 0:02:15.160 of a trans humanist bill of rights, and it included 0:02:15.200 --> 0:02:19.520 things like the right to extend one's life and stuff 0:02:19.560 --> 0:02:23.840 like that. But anyway, so, uh so, Istabon has a 0:02:23.880 --> 0:02:27.720 thesis that he published in a Vice Motherboard article. Uh 0:02:28.000 --> 0:02:30.440 that was I guess a year or two ago, I think, 0:02:30.960 --> 0:02:35.240 and he he said, you know what artificial general intelligence 0:02:35.400 --> 0:02:38.240 is not going to arise in the lab of a 0:02:38.280 --> 0:02:42.640 tech company like Google, or a university research program or 0:02:42.720 --> 0:02:46.000 some genius kids garage. A g I is going to 0:02:46.040 --> 0:02:50.160 arise through the work of state actors defending their interests 0:02:50.840 --> 0:02:56.079 and as so, speaking of coming superhuman artificial intelligence, Zoltan 0:02:56.240 --> 0:02:59.959 rights quote, politicians and military commanders around the world will 0:03:00.120 --> 0:03:03.600 want this super intelligent machine mind for their countries and 0:03:03.639 --> 0:03:09.000 defensive forces, and they'll want it exclusively. Using AI's potential 0:03:09.080 --> 0:03:13.280 power and might for national security strategy is more than obvious. 0:03:13.520 --> 0:03:17.800 It's essential to retain leadership in the future world. Inevitably, 0:03:17.840 --> 0:03:22.040 a worldwide AI arms race is set to begin. So 0:03:22.040 --> 0:03:24.840 so here's the thesis that we can explore for the 0:03:24.880 --> 0:03:29.080 rest of today's episode. Now, Ralph, the face of it, 0:03:29.120 --> 0:03:33.600 I think at least part of Sultan's argument is is wrong. Yeah, 0:03:34.360 --> 0:03:37.560 I don't necessarily think the AI arms race part is wrong, 0:03:37.840 --> 0:03:42.640 And I don't think necessarily that nations wouldn't want exclusive 0:03:42.800 --> 0:03:45.800 use of some sort of super intelligent entity to work 0:03:45.840 --> 0:03:49.800 on their behalf. I think that's apparent right, Like it 0:03:49.840 --> 0:03:52.680 would be crazy to suggest otherwise that if there was 0:03:53.080 --> 0:03:56.800 any given nation were to have the opportunity to have 0:03:56.880 --> 0:04:00.840 a super intelligent entity on their side. It would be 0:04:00.880 --> 0:04:04.920 insane to say, now we're good, go help someone else 0:04:05.040 --> 0:04:07.840 that then what's the part you disagree with? I disagree 0:04:07.880 --> 0:04:10.000 with his statement that says that AI is not going 0:04:10.040 --> 0:04:12.040 to show up in the lab of a tech company, 0:04:12.120 --> 0:04:15.320 but instead will become the will arise because of the 0:04:15.320 --> 0:04:18.080 the work of state actors. Yeah, okay, so, and I 0:04:18.120 --> 0:04:20.200 know that's not the only thing we're going to disagree with, 0:04:20.320 --> 0:04:22.159 but that's the first thing here. Yeah, that's the first 0:04:22.160 --> 0:04:25.480 thing I disagree with, because I don't know if you've 0:04:25.520 --> 0:04:29.440 ever worked for any government agency, Joe, have you? Have 0:04:29.520 --> 0:04:33.279 you ever worked for I've worked for a university's state universe. 0:04:33.400 --> 0:04:37.119 That's close. Yeah, Because all right, so you know about bureaucracy, 0:04:37.440 --> 0:04:40.760 You know about the barriers that are there and for progress, 0:04:40.800 --> 0:04:43.600 because well I know about bureaucracy and barriers to progress 0:04:43.640 --> 0:04:46.760 in private industry. Sure, sure, it's just in government, and 0:04:46.839 --> 0:04:49.680 I think it tends to be on a on a 0:04:49.760 --> 0:04:54.839 level that's exacerbated prodigious, right, And I mean, let's be fair. Okay, 0:04:54.839 --> 0:04:59.280 So bureaucracy exists for a reason. Bureaucracy exists so that 0:04:59.720 --> 0:05:04.960 p did tasks have a streamlined process to go through, right, 0:05:05.600 --> 0:05:08.680 so that that way you have designed a system that 0:05:08.839 --> 0:05:12.679 is really good at handling very specific things. The problem 0:05:12.760 --> 0:05:17.320 is anything that's outside of that specificity doesn't fit the system, 0:05:17.520 --> 0:05:21.760 and then getting any progress on that is laborious at 0:05:21.800 --> 0:05:28.080 best or possibly impossible, depending upon the degree. Now, I 0:05:28.080 --> 0:05:30.839 would say that companies are far more nimble and have 0:05:30.960 --> 0:05:35.000 a greater incentive to invest in research and development of 0:05:35.080 --> 0:05:39.520 artificial intelligence on a scale that dwarfs anything that any 0:05:39.600 --> 0:05:42.760 government is able to do, unless you're talking about a 0:05:42.920 --> 0:05:48.200 massive government that just decides to turn its full attention 0:05:48.360 --> 0:05:52.920 on this problem. So king yes, yeah, some sort of 0:05:53.880 --> 0:05:57.640 very authoritarian approach saying all right, we're just turning all 0:05:57.680 --> 0:06:00.280 of our like if you're playing the game Civilization, you 0:06:00.320 --> 0:06:03.080 switch all your cities so that they're just producing science 0:06:03.520 --> 0:06:05.839 that kind of thing. Where uh, you know, here in 0:06:05.839 --> 0:06:09.600 the United States, for example, yes, we have organizations like 0:06:09.720 --> 0:06:13.760 DARPA that's part of the Department of Defense. They are 0:06:13.920 --> 0:06:18.919 very heavily into administering projects that guide research and development 0:06:18.920 --> 0:06:21.880 in AI and other areas of technology. They're not the 0:06:21.920 --> 0:06:25.080 ones who do it themselves. Private companies or even publicly 0:06:25.120 --> 0:06:29.280 traded companies or research organizations do the work. So I 0:06:29.360 --> 0:06:32.479 think maybe state actors will play a part in the 0:06:32.520 --> 0:06:35.480 sense that there might be some funding. But I would 0:06:35.600 --> 0:06:38.760 argue that we're far more likely to see the emergence 0:06:39.000 --> 0:06:44.600 of something approaching general artificial intelligence from one of these companies, 0:06:45.040 --> 0:06:50.680 not from like a specific state sponsored program, because the 0:06:50.680 --> 0:06:54.719 companies are more nimble. They're able to respond and change 0:06:55.279 --> 0:06:59.080 their course of action much more quickly than any government 0:06:59.080 --> 0:07:03.600 agency can. There's a lot of inertia and momentum involved 0:07:03.600 --> 0:07:06.880 in government agencies, and trying to change course is very 0:07:06.880 --> 0:07:08.800 hard to do. You know. I would say, if you 0:07:08.880 --> 0:07:13.440 are a government and you're interested in funding AI research 0:07:13.560 --> 0:07:15.880 to win this AI arms race, to beat all the 0:07:15.880 --> 0:07:21.040 other countries and international competitors to beat, to establish this 0:07:21.200 --> 0:07:23.600 a g I that you can call upon to do 0:07:23.680 --> 0:07:27.600 your bidding, I think that the best strategy to go 0:07:27.680 --> 0:07:30.480 about that would probably be to set some kind of 0:07:30.600 --> 0:07:34.560 huge prize for it. Uh and and governments have done 0:07:34.560 --> 0:07:38.080 things like this before. Do yeah, research prizes, So some 0:07:38.200 --> 0:07:42.080 gigantic pool of money, you say, first person to produce 0:07:42.120 --> 0:07:45.560 a computer that can meet these five requirements gets all 0:07:45.600 --> 0:07:49.080 of this, yes and uh and so again, then you're 0:07:49.160 --> 0:07:52.880 still talking on indirect kind of influence, right, like that 0:07:52.960 --> 0:07:56.040 you've you've created the incentive, but it's not the agency 0:07:56.160 --> 0:07:58.880 that actually creates the a g I that has been 0:07:59.040 --> 0:08:02.520 the Resulten has been argueing for. Also in another article 0:08:02.560 --> 0:08:06.160 at Sultan wrote there's a bit where uh he he says, well, 0:08:06.200 --> 0:08:09.520 this would be unlike the space industry, where we see 0:08:09.880 --> 0:08:13.080 international cooperation like on the International Space Station, which to 0:08:13.120 --> 0:08:16.600 me seems to be very convenient in the fact that 0:08:16.640 --> 0:08:20.480 there he's ignoring the origin of the space race, like 0:08:20.520 --> 0:08:23.760 he's saying, AI would be different from the space race 0:08:23.800 --> 0:08:26.360 because with the space race, or rather the space industry, 0:08:26.480 --> 0:08:31.000 because the space industry, we see this relatively peaceful cooperation 0:08:31.120 --> 0:08:35.719 between nations to advance our understanding of science and technology. Right, 0:08:37.160 --> 0:08:40.160 But the space race started as a branch of the 0:08:40.160 --> 0:08:43.239 Cold War. I mean it was, it was a product 0:08:43.360 --> 0:08:46.280 of the Cold War. If it weren't for this this 0:08:46.679 --> 0:08:49.440 our space program, could could you argue that it's essentially 0:08:49.480 --> 0:08:53.040 something that was like hastily cobbled together after sput Nick 0:08:53.520 --> 0:08:56.600 because they were like, oh crap, well yeah, I mean, 0:08:56.640 --> 0:08:59.839 and spot Nick did exactly what the Russians wanted it 0:08:59.880 --> 0:09:02.480 to do, which was to indicate to the United states, Hey, 0:09:02.520 --> 0:09:04.760 we can build a rocket that can reach you, right, 0:09:04.840 --> 0:09:08.920 that was really the Soviet Union's primary goal. Now, the 0:09:08.920 --> 0:09:11.360 people who worked on the spot Nik project, they had 0:09:11.400 --> 0:09:15.079 their other like their their own individual motivations and their 0:09:15.080 --> 0:09:18.480 own individual goals for that project, but the state goal 0:09:18.960 --> 0:09:21.800 was to say, we can launch a rocket that can 0:09:21.840 --> 0:09:26.160 reach you. You America, And then America said, well, we 0:09:26.240 --> 0:09:28.760 have to make sure that we have the same capability. 0:09:29.240 --> 0:09:32.600 Um and they funded it in a way where it 0:09:32.720 --> 0:09:35.280 was in the name of science, and there were legitimate 0:09:35.320 --> 0:09:39.080 scientists doing legitimate science under this program. But it wouldn't 0:09:39.120 --> 0:09:42.160 have existed without the Cold War, right, without this this 0:09:42.240 --> 0:09:48.679 competition between two massive world powers. And so really I 0:09:48.720 --> 0:09:51.240 would argue that the reason why the space industry is 0:09:51.280 --> 0:09:54.520 the way it is is because this history that predates 0:09:54.559 --> 0:09:59.839 it of of competition and kind of uh well again 0:10:00.679 --> 0:10:02.680 a branch of that Cold War, it was, it was 0:10:03.679 --> 0:10:07.920 acts of passive aggression, I guess, uh so, I think 0:10:07.920 --> 0:10:10.560 an AI arms race. I agree with this point that 0:10:10.600 --> 0:10:13.520 an AI arms race would be uh something we would 0:10:13.520 --> 0:10:17.640 expect should we reach that level of AI sophistication. And 0:10:17.679 --> 0:10:19.880 you could even argue that we're seeing an AI arms 0:10:19.960 --> 0:10:22.359 race right now. It's just we're seeing it in the 0:10:22.400 --> 0:10:26.080 context of narrow AI as opposed to general AI. But 0:10:26.480 --> 0:10:30.079 I a lot of the premises he sets up to 0:10:30.200 --> 0:10:33.280 support that I disagree with. I think that that those 0:10:33.320 --> 0:10:35.760 would need to be revisited. I do agree with his 0:10:36.960 --> 0:10:40.520 end conclusion, but his argument is not I don't think 0:10:40.520 --> 0:10:43.760 it's supportable like the actual premises of his argument. That 0:10:43.880 --> 0:10:47.360 is okay, Well, what are some of the basic arguments 0:10:47.360 --> 0:10:51.760 than about whether the world is about to be entering 0:10:52.320 --> 0:10:57.160 an artificial intelligence arms race? Well, like Sultan says, he says, 0:10:57.200 --> 0:11:02.720 an AI a really useful AI. So the furthest extension 0:11:02.720 --> 0:11:05.120 we would put to that is the super intelligent general 0:11:05.320 --> 0:11:08.720 artificial intelligence. Although I argue that's not necessary for it 0:11:08.760 --> 0:11:11.040 to be an issue. Uh. He says, it would be 0:11:11.080 --> 0:11:13.719 such a valuable tool that no one would be able 0:11:13.800 --> 0:11:17.680 to ignore that possibility, like to to to say, we're 0:11:17.760 --> 0:11:20.160 not going to go down that road. Well, I'd agree 0:11:20.160 --> 0:11:23.160 that AI is a tool too powerful to ignore. I mean, 0:11:23.600 --> 0:11:28.720 in general, intelligence is the most valuable thing we have. 0:11:28.960 --> 0:11:32.600 Intelligence is the thing that you know, not just that 0:11:32.760 --> 0:11:35.480 empowers us, that makes human life worth living, I mean, 0:11:35.600 --> 0:11:40.760 and and increasing your share of intelligence capability is probably 0:11:40.880 --> 0:11:44.080 the most important thing you could do to ensure your 0:11:44.200 --> 0:11:49.480 future success. Right, and now he makes the argument that 0:11:49.720 --> 0:11:53.880 whichever nation out there created the creates the first super 0:11:53.880 --> 0:11:57.720 intelligent a g I wins the game. Like, imagine that 0:11:57.720 --> 0:12:03.440 that geo political uh influences and forces are all reduced 0:12:03.480 --> 0:12:06.880 to a board game. Whoever hits that last square is 0:12:06.880 --> 0:12:09.840 the one who hit first, is the one who creates 0:12:09.960 --> 0:12:11.880 super intelligent a g I, and they win the whole 0:12:11.920 --> 0:12:16.480 thing automatically, hands down. Um. I don't necessarily believe that, 0:12:16.559 --> 0:12:19.640 because I think that access to a super intelligent a 0:12:19.720 --> 0:12:22.920 g I doesn't immediately solve other issues that a nation 0:12:23.040 --> 0:12:28.080 must address, at least not not so quickly that other 0:12:28.160 --> 0:12:31.640 nations can't also react to it. Yeah, I I think 0:12:31.679 --> 0:12:34.160 I'd agree with you there. I came up with a 0:12:34.240 --> 0:12:37.360 little thought experiment to sort of service an analogy. So 0:12:37.480 --> 0:12:40.200 imagine you've got two ships trying to win a race 0:12:40.280 --> 0:12:44.480 across the Atlantic Ocean. One ship is a nineteenth century 0:12:44.559 --> 0:12:48.400 style steamship, is capable of about eight knots, and the 0:12:48.480 --> 0:12:51.800 captain is assisted by a team of the world's most 0:12:51.880 --> 0:12:55.440 brilliant scientists with access to powerful supercomputers and all the 0:12:55.440 --> 0:12:58.640 tools they could want. The other ship is a modern 0:12:58.679 --> 0:13:02.800 ocean liner powered by geared steam turbines capable of thirty knots, 0:13:02.800 --> 0:13:06.319 but it is captained by a rather dim guy named Todd, 0:13:06.440 --> 0:13:13.560 who has no scientists or supercomputers at his disposal at all. Todd, Now, unfortunately, 0:13:13.600 --> 0:13:17.720 I think the fact is given what I've stated, Todd 0:13:17.800 --> 0:13:20.559 is still going to win the race, right unless unless 0:13:20.600 --> 0:13:25.200 Todd is so dumb as to cause damage to his ship. Yes, then, 0:13:25.559 --> 0:13:28.000 assuming that Todd is is at least of the level 0:13:28.000 --> 0:13:32.760 of intelligence necessary to follow a specific route, Yeah, it's so. Yeah, 0:13:32.800 --> 0:13:35.240 it's good to have very smart people at your disposal 0:13:35.320 --> 0:13:38.000 to help optimize your chances of winning. But if you're 0:13:38.000 --> 0:13:40.680 stuck with a ship that's more than three times slower 0:13:41.360 --> 0:13:43.280 and and that's how you start the race, you're just 0:13:43.320 --> 0:13:47.120 probably not gonna win. But I would say maybe if 0:13:47.120 --> 0:13:49.320 the captain of the first ship is able to consult 0:13:49.360 --> 0:13:52.280 with the scientists and the supercomputers over a period of 0:13:52.320 --> 0:13:55.679 several years leading up to the race, well then you 0:13:55.679 --> 0:13:57.760 you could maybe go in with a clear advantage and 0:13:57.800 --> 0:14:01.319 win it. But then again, if if dim Todd knows this, 0:14:01.520 --> 0:14:04.240 he may get some scientists and supercomputers of his own 0:14:04.640 --> 0:14:06.680 and then you've got a right, you've got an arms 0:14:06.720 --> 0:14:09.280 race on your hands. But I would say that just 0:14:09.400 --> 0:14:14.480 having this intelligence advantage does not automatically mean you immediately 0:14:14.559 --> 0:14:17.440 become the runaway winner for all time, right though. It 0:14:17.520 --> 0:14:20.640 though it is a very strong advantage. And ce Zoltan's 0:14:20.680 --> 0:14:24.400 point is that, or the argument he makes is that 0:14:25.400 --> 0:14:29.520 whichever nation develops at first will have a dominant hand 0:14:29.720 --> 0:14:32.720 from that point forward, will become the dominant force on 0:14:32.760 --> 0:14:34.800 the face of the planet. Yeah, I know. He's also 0:14:34.880 --> 0:14:37.360 of the opinion that I think, if you have the power, 0:14:37.760 --> 0:14:39.960 so if you have a g I on your side, 0:14:40.480 --> 0:14:43.840 you have the power to shut down and prevent other 0:14:43.960 --> 0:14:46.920 nations from achieving a g I. And I think unless 0:14:46.920 --> 0:14:51.040 you're willing to take incredibly extreme measures, that's not necessarily 0:14:51.040 --> 0:14:54.360 the case either. It also presupposes that that a g 0:14:54.600 --> 0:14:59.280 I is going to have those capabilities, and it may 0:14:59.320 --> 0:15:03.080 be that it can accomplish things that we cannot easily 0:15:03.120 --> 0:15:05.360 do on our own, but doesn't necessarily mean it can 0:15:05.360 --> 0:15:10.040 magically accomplish anything, right Like I there's I think he's 0:15:10.080 --> 0:15:13.320 giving it so much credit. And to be fair, we 0:15:13.360 --> 0:15:17.320 don't know, right we we don't have anything that's super intelligent, 0:15:17.440 --> 0:15:20.000 because we've not been capable of creating such a thing. 0:15:20.440 --> 0:15:24.000 And maybe he's right, Maybe a super intelligent entity would 0:15:24.040 --> 0:15:26.720 be able to accomplish things that most of us would 0:15:26.720 --> 0:15:29.960 think of as being, if not impossible, impractical because it 0:15:29.960 --> 0:15:34.200 would be so difficult to do. But I agree with you. 0:15:34.240 --> 0:15:36.040 I think that it would be something that would give 0:15:36.040 --> 0:15:40.720 a country a distinct advantage, but not guarantee what Zoltan 0:15:40.880 --> 0:15:44.000 seems to think of as victory. Uh. He he seems 0:15:44.040 --> 0:15:47.240 to to, at least in his arguments. I think that 0:15:48.640 --> 0:15:54.120 what we're heading toward is a world that's united because 0:15:54.200 --> 0:15:58.280 whichever country comes up with a super intelligent a g 0:15:58.440 --> 0:16:01.160 I will be able to subjugate all other countries do 0:16:01.280 --> 0:16:03.760 its will. And he thinks, gosh, darn it, the good 0:16:03.760 --> 0:16:08.720 old US of A should be the country to do that. Ah. 0:16:08.800 --> 0:16:13.080 There's some elements to Sultan's philosophy that I find particularly troubling. 0:16:13.160 --> 0:16:15.600 Might be might be too strong a word. Well, I 0:16:15.600 --> 0:16:18.560 mean to defend I I don't think I agree with 0:16:18.600 --> 0:16:21.080 his philosophy either, but to defend what he says, I mean, 0:16:21.160 --> 0:16:24.480 he he does say at least what he has in 0:16:24.600 --> 0:16:28.320 mind is a sort of benevolent hegemon e. It's not 0:16:28.360 --> 0:16:30.480 like we're going to subjugate you and oppress you and 0:16:30.520 --> 0:16:32.040 do what we tell you. I think what he has 0:16:32.080 --> 0:16:33.800 in mind is that, you know, it's going to be 0:16:33.880 --> 0:16:38.120 the Federation from Star Trek. Everybody's happy, and you know, 0:16:38.200 --> 0:16:40.320 it's a utopian kind of I think that's what he 0:16:40.360 --> 0:16:44.320 sees if the United States is the entity that creates this. 0:16:44.440 --> 0:16:48.200 But I think he also imagines a world that would 0:16:48.200 --> 0:16:51.760 be in the realm of the dystopian future should some 0:16:51.920 --> 0:16:55.120 other country do it, like, for example, China or Russia. 0:16:55.480 --> 0:16:58.320 Not that he specifically names these countries, but that would 0:16:58.360 --> 0:17:00.240 be example, or North Korea. I think he had actually 0:17:00.280 --> 0:17:02.840 does in North Korea. No, I will admit I can 0:17:02.880 --> 0:17:06.680 imagine if North Korea were able to develop an artificial 0:17:06.720 --> 0:17:09.640 general intelligence long before anyone else. I think that could 0:17:09.680 --> 0:17:13.760 have disastrous consequences for the planet. Absolutely. But I think 0:17:13.800 --> 0:17:18.040 again saying that in those scenarios, you don't get that 0:17:18.160 --> 0:17:21.680 Starfleet future, you also get some other science fiction future 0:17:21.720 --> 0:17:24.880 that's probably written by Burgess or something. I also don't 0:17:24.920 --> 0:17:28.359 think that's very likely though. Yeah. Uh, Now we have 0:17:28.480 --> 0:17:33.440 other people that have weighed in on this issue, who 0:17:34.040 --> 0:17:37.760 have some really interesting points to Yeah. One thing we 0:17:37.800 --> 0:17:41.040 read is by Anya Cosperson, writing for the World Policy 0:17:41.119 --> 0:17:44.960 Journal blog about the idea of an AI arms race, 0:17:45.000 --> 0:17:47.200 and what does she say, Jonathan. She says that, first 0:17:47.240 --> 0:17:52.160 of all, AI is it's undeniably intrinsically valuable. And Yeah, 0:17:52.240 --> 0:17:54.040 I think that. I think there is no way to 0:17:54.080 --> 0:17:56.520 dispute that. I think we both agree that AI is 0:17:56.600 --> 0:18:00.639 definitely valuable. Um, even in even at the extent that 0:18:00.680 --> 0:18:04.040 we have it right now, the the narrow AI that 0:18:04.200 --> 0:18:08.880 is our reality, we see that it's valuable because it's 0:18:08.920 --> 0:18:12.159 doing stuff for us. I mean that that kind of 0:18:12.200 --> 0:18:17.199 defines value. Uh. So there's an incentive for countries to 0:18:17.280 --> 0:18:22.600 pursue the development of AI. I think that's undeniable too. Now, 0:18:22.640 --> 0:18:26.080 even if those countries agree that weaponized AI is unethical, 0:18:26.560 --> 0:18:29.520 that's not enough of a deterrent, she says, to make 0:18:29.560 --> 0:18:33.119 sure that this does not, you know, blossom into a 0:18:33.200 --> 0:18:37.160 weaponized AI future, because Yeah, there's always gonna be rogue 0:18:37.200 --> 0:18:41.040 operators that may not be state sponsored. You may have 0:18:41.320 --> 0:18:45.040 independent rogue operators that are willing to either develop or 0:18:45.720 --> 0:18:49.520 perhaps more likely appropriate AI technology and then convert it 0:18:49.560 --> 0:18:53.879 for weaponized purposes, which creates pressure on all nations to 0:18:53.960 --> 0:18:58.800 consider developing weaponized AI as uh, just as a preventive measure. 0:18:59.440 --> 0:19:02.240 So in other Mr President, we cannot allow an AI 0:19:02.320 --> 0:19:06.399 gap exactly exactly. Uh. There are a lot of doctor 0:19:06.480 --> 0:19:10.600 Strange love moments I had while researching this this particular topic. 0:19:11.280 --> 0:19:14.080 You can't fight here. This is the war room. So 0:19:14.720 --> 0:19:17.040 one of the examples that you could think of is 0:19:17.080 --> 0:19:19.879 that the United Nations agrees that weaponized DAI is a 0:19:19.920 --> 0:19:23.480 bad idea. This is this is a hYP hypothetical situation. 0:19:23.840 --> 0:19:25.880 I actually had more to say about this in a bit. Yeah, 0:19:25.880 --> 0:19:28.000 well we'll we'll cover it in greater detail in a second. 0:19:28.080 --> 0:19:31.919 So in this hypothetical situation, the U n says, all right, uh, 0:19:32.119 --> 0:19:35.879 we as an international community agree that weaponized DAI is 0:19:35.880 --> 0:19:38.640 a bad idea. It's gonna lead us down a destructive road. 0:19:39.119 --> 0:19:42.000 Once we head down there, there's no turning back, and 0:19:42.040 --> 0:19:45.879 the outcome is going to be negative. There's not a 0:19:45.920 --> 0:19:49.080 way of envisioning this future where there's like a net 0:19:49.080 --> 0:19:54.040 positive outcomes. So we're gonna agree no weaponized DAI. But 0:19:54.080 --> 0:19:57.159 then you've got those rogue operators like a terrorist cell 0:19:57.440 --> 0:20:00.879 or organized crime that see this as an opportunity to 0:20:01.000 --> 0:20:04.960 use AI in a weaponized way, and they don't. They're 0:20:04.960 --> 0:20:08.399 not beholden to this international agreement, they have no stake 0:20:08.440 --> 0:20:11.359 in that, right, They're they're not getting they're not worried 0:20:11.400 --> 0:20:15.000 about international sanctions from the u N. They're trying to 0:20:15.040 --> 0:20:19.080 accomplish whatever goals their particular organization has and they're using 0:20:19.119 --> 0:20:21.520 whatever tools are at their disposal to do so. Yeah, 0:20:21.600 --> 0:20:25.320 I mean you're invoking here the concept that and this 0:20:25.400 --> 0:20:28.520 is true generally even peace loving people who are not 0:20:28.640 --> 0:20:31.639 looking to start a fight don't tend to opt for 0:20:31.800 --> 0:20:35.800 unilateral disarmament. Yeah, because because if you were to do 0:20:35.840 --> 0:20:39.120 that and there's that one person who hasn't agreed, then 0:20:39.119 --> 0:20:41.520 that one person just runs rampant over everybody else. Right, 0:20:41.880 --> 0:20:44.679 Like if you if you had a group of kids, 0:20:45.080 --> 0:20:47.160 let's go Lord of the Flies. You have a group 0:20:47.200 --> 0:20:50.720 of kids, and each kid is given a whiffleball bat, 0:20:51.760 --> 0:20:54.040 and it's just to keep this from being like too 0:20:54.119 --> 0:20:57.400 violent for for even for a hypothetical situation. Each kid 0:20:57.400 --> 0:21:00.800 has a whiffleball bat, and you tell all the kids, like, well, 0:21:01.080 --> 0:21:04.359 you can either have that whiffleball bat and wail on 0:21:04.440 --> 0:21:06.920 your classmates, or you can put the whiffleball bat down 0:21:07.160 --> 0:21:10.280 and we'll have a fun activity. But you all need 0:21:10.320 --> 0:21:12.520 to put the whiffleball bat down in order for that 0:21:12.560 --> 0:21:15.280 to happen. And one by one they start dropping their 0:21:15.320 --> 0:21:17.400 whiffleball bats until there's that one bully in the back. 0:21:17.440 --> 0:21:19.399 It's like, no, I know what my fun activity is, 0:21:19.560 --> 0:21:22.560 hit all the kids who don't have whiffleball bats. That's 0:21:22.600 --> 0:21:25.439 the same sort of argument here and up. So you know, 0:21:25.480 --> 0:21:27.639 I made a little funny note in here saying, next thing, 0:21:27.680 --> 0:21:30.760 you know, we've got gangster drones shaking down local businesses 0:21:30.800 --> 0:21:34.880 for extortion money, Like you know, it's a you gotta 0:21:35.000 --> 0:21:37.000 you got a nice pizza joint here, and be a 0:21:37.000 --> 0:21:40.240 shame as someone I don't know assimilated it into what's 0:21:40.240 --> 0:21:46.760 growing technological layout, you know. So so you know, I 0:21:46.800 --> 0:21:51.280 think that's a legitimate argument, and I agree very much 0:21:51.280 --> 0:21:56.159 with those those assertions that if in fact, you know 0:21:56.600 --> 0:21:59.840 there's a possibility to go down this, then someone, someone, 0:22:00.320 --> 0:22:02.879 whether state sponsor or not, is going to do it. 0:22:03.240 --> 0:22:06.560 I would also argue we're already at a point where 0:22:06.600 --> 0:22:10.199 this can happen. You don't need a g I for 0:22:10.240 --> 0:22:14.400 this to be a problem. Narrow AI is enough exactly, 0:22:14.480 --> 0:22:18.840 So you don't need a general intelligence that can you know, 0:22:19.160 --> 0:22:23.359 be super intelligent and produce the next next generation of 0:22:23.400 --> 0:22:26.879 technology and innovations and everything. You just need like a 0:22:26.960 --> 0:22:31.240 really good weapon that's smart. Yeah. Or even if you're 0:22:31.680 --> 0:22:35.320 going to go the semi autonomous route right like, you 0:22:35.359 --> 0:22:38.800 could argue we're there with drones. Were there, we're using 0:22:38.840 --> 0:22:45.639 technology that removes human operatives from conflict and replaces them 0:22:45.720 --> 0:22:48.480 with robotic ones. We'll have a lot more to say 0:22:48.480 --> 0:22:51.240 about about that concept a little bit further too, because 0:22:51.240 --> 0:22:54.680 that that ties into one of the other big concerns 0:22:54.720 --> 0:22:59.080 about weaponized AI in general. Sure, okay, so let's uh, 0:22:59.160 --> 0:23:02.600 let's speculate little more. Sure, if there is an AI 0:23:02.800 --> 0:23:05.399 arms race, what's it gonna look like. What are some 0:23:05.480 --> 0:23:09.520 scenarios that could actually happen? Okay, So we have the 0:23:09.560 --> 0:23:16.080 state sponsored approach where countries decide not to essue weabinized AI. 0:23:16.720 --> 0:23:20.959 Whether that ends up being a direct implementation of like 0:23:21.040 --> 0:23:25.280 a robotic soldier or AI that's assisting in some military 0:23:25.280 --> 0:23:29.200 capacity for strategical or tactical planning, it doesn't really matter. 0:23:29.280 --> 0:23:32.280 You could have topped down AI or bottom up AI exactly, 0:23:32.400 --> 0:23:34.560 or both you know, you could have some integrated approach 0:23:34.560 --> 0:23:37.440 that's both. Uh. So one scenario is that you have 0:23:37.560 --> 0:23:42.960 established nations like the United States, Russia, China developing weaponized 0:23:43.320 --> 0:23:46.920 AI like webinized robots is a good example. So there's 0:23:47.160 --> 0:23:49.760 still some they're already examples like drones that are kind 0:23:49.800 --> 0:23:53.680 of in this category, whether they are autonomous or semi 0:23:53.760 --> 0:23:58.080 autonomous or remote controlled. Uh, there are various types. Obviously 0:23:58.119 --> 0:24:00.240 remote controlled you wouldn't call it AI because it's under 0:24:00.240 --> 0:24:03.240 the operation of a of a human, but semi autonomous 0:24:03.280 --> 0:24:05.439 or autonomous would fall into the AI category, even the 0:24:05.520 --> 0:24:10.080 narrow AI band. So these future robots, uh that these 0:24:10.119 --> 0:24:13.320 hypothetical future robots that are weaponized AI, they take the 0:24:13.320 --> 0:24:17.840 place of our soldiers, or they augment the presence of soldiers. 0:24:18.320 --> 0:24:20.600