Annie Jacobsen Explains the Terrifying Side of Artificial Intelligence (YouTube Video Transcript)

(00:00:00) Your YouTube transcript will appear here (00:00:00) I'm conflicted. What do you do? I think (00:00:02) we're at a new crossroads. (00:00:05) >> Keep going. I'm interested what you have (00:00:07) to say. (00:00:07) >> Well, I mean, there's a lot of fear of (00:00:08) what AI could develop into and China's (00:00:11) not going to stop. We stop, then we put (00:00:13) ourselves at a disadvantage. But if none (00:00:15) of us stop, we put the entire human (00:00:17) species at a risk of becoming (00:00:18) irrelevant. (00:00:19) >> You have biological warfare threats that (00:00:22) become more existentially threatening (00:00:24) with the introduction of AI, perhaps (00:00:26) more so than with the nuclear weapons. I (00:00:28) think we're at a new crossroads (00:00:33) that's completely different but maybe (00:00:35) has some similarities with AI. (00:00:37) >> Okay. (00:00:38) >> Do you (00:00:40) >> keep going? I'm interested what you have (00:00:42) to say. (00:00:42) >> Well, I mean, there's a lot of fear of (00:00:43) what AI could develop into and (00:00:45) basically, you know, gets to the point (00:00:47) where it makes humanity completely (00:00:50) irrelevant. And so, (00:00:54) it's it's it's it's a dangerous game (00:00:56) we're playing. I don't think anybody (00:00:57) really knows the extent of what we might (00:01:01) experience if this keeps keeps on, you (00:01:04) know, with the brain chips with with (00:01:05) everything. (00:01:06) >> And so it's another it's (00:01:11) >> once it's out of the bag, it's not going (00:01:13) back in. And I feel like that's where (00:01:15) we're at with AI is where we were with (00:01:17) nuclear war. (00:01:19) >> It's the Pandora's box idea. Yes. (00:01:22) which kind of begs the question, (00:01:26) okay, so how about doing something about (00:01:28) it? And I know we all have busy lives (00:01:30) and that is part of the, you know, no (00:01:34) one can stop what they're doing and (00:01:35) suddenly uh become an expert on (00:01:38) existential threats per se. So we talk (00:01:41) about it and we (00:01:44) you know pontificate what can be done (00:01:48) but we should remember that there are (00:01:51) powers that be that are paid to deal (00:01:53) with these issues in our own government. (00:01:55) Mhm. (00:01:56) >> And also the days are over where you (00:01:59) could just trust (00:02:01) the government (00:02:04) to be doing (00:02:06) the how do I say that right? (00:02:10) >> Yeah. I know where you're going. We (00:02:12) should be able to trust the government (00:02:13) to have our to be doing things in our (00:02:15) best interest. (00:02:16) >> Yeah. Right. Like if you watch the (00:02:18) propaganda films of the 50s having to do (00:02:20) with nuclear war, right? I mean, and you (00:02:24) see like a housewife with a tiny waist (00:02:26) and you know, curlers maybe even making (00:02:28) pancakes and then like a siren goes off (00:02:31) and then Jimmy come quick. And they they (00:02:34) duck and cover and that's going to (00:02:36) protect you against a nuclear bomb and (00:02:37) everybody went okay. I mean, those days (00:02:42) are over, but then you kind of I think (00:02:46) the point you're raising is what kind of (00:02:49) like (00:02:51) version of that are we dealing with (00:02:53) today? (00:02:55) >> I I think it's AI. I mean, it's (00:02:57) happening right now. You know, (00:03:00) Xi Jinping in China says the first the (00:03:03) first country that masters (00:03:06) AI will achieve global domination. And (00:03:10) so now you have now you have all these (00:03:14) people over here that are worried about (00:03:16) I mean I'm conflicted. What do you do? (00:03:19) China's not going to stop, (00:03:20) >> right? (00:03:20) >> They're not going to stop. And (00:03:24) so if we stop, then we put ourselves at (00:03:27) a disadvantage. But if none of us stop, (00:03:30) we put the entire human species at a at (00:03:33) a at the risk of becoming irrelevant, (00:03:37) you know, and and does that make sense? (00:03:40) >> Of course, it makes sense and it's (00:03:42) absolutely on point. And you know, you (00:03:44) can also throw into that mix (00:03:48) biology (00:03:50) because, (00:03:52) okay, so here's how I would tie that (00:03:53) together, right? Because you have (00:03:55) biological warfare threats (00:03:58) that become (00:04:02) more threatening, more existentially (00:04:04) threatening with the introduction of AI. (00:04:06) I believe perhaps more so than with the (00:04:08) nuclear weapons. Right. One of the areas (00:04:12) I'm going to try to hold this thought (00:04:13) together, but it it ties. (00:04:16) So, (00:04:18) you might say, you might say nuclear (00:04:20) weapons could, you know, AI could get (00:04:22) hold of nuclear weapons. (00:04:26) Well, maybe. And this is where I'm (00:04:28) either (00:04:30) informed or inaccurate. I don't know. (00:04:33) Okay. (00:04:36) What I do know is that from from (00:04:38) interviewing people in cyber command is (00:04:40) that our nuclear weapons are (00:04:43) surprisingly analog, (00:04:45) meaning they are not digital. (00:04:48) >> Okay. So, for example, I learned in (00:04:51) reporting the book that our sublistic (00:04:53) missiles guide to the targets by star (00:04:57) sighting. Sean, a little panel opens (00:05:02) and they use the stars to guide to the (00:05:04) target. There are other systems in place (00:05:06) and this stuff is very classified but (00:05:08) what is leading is like this ancient (00:05:10) technology that like our huntergatherer (00:05:13) ancestors used. Okay. Um so nuclear (00:05:19) weapons because they happened before the (00:05:21) advent of the digital age there has been (00:05:24) a concerted effort to make sure they (00:05:28) remain analog so that they can't be (00:05:31) hacked. (00:05:31) >> Okay. (00:05:32) >> Okay. And these are assurances that I (00:05:34) have gotten from Cyber Command. You're (00:05:37) just taking somebody at your at their (00:05:38) word at that point because the (00:05:40) documentation is not declassified. (00:05:43) >> So hold that thought. Then you have this (00:05:48) idea that (00:05:51) biological weapons used to exist. We (00:05:53) used to have we had a program about (00:05:55) biological weapons. I wrote about it. We (00:05:56) hired the Nazi scientists. They built up (00:05:59) our biological warfare program. and we (00:06:02) used to have an arsenal and then Nixon (00:06:07) made them illegal. So all of the (00:06:09) biological weapons were destroyed. (00:06:12) We found out Russia was cheating and (00:06:14) they I mean rat hole upon rat hole. Um (00:06:19) so biological weapons are no more which (00:06:21) exist sort of that's the reason that (00:06:24) nuclear disarmament people say we don't (00:06:25) need an nuclear arsenal to keep us safe (00:06:29) because we we were able to say we don't (00:06:32) need biological weapons to keep us safe (00:06:34) biological weapons have become taboo (00:06:37) we need so the disarmament people will (00:06:39) say nuclear weapons should be taboo now (00:06:42) you take AI okay what you're saying (00:06:45) which is really significant to think (00:06:47) about is how does AI fit into the mix? (00:06:51) If there is indeed a giant gap on (00:06:54) purpose between AI being able to access (00:06:57) nuclear weapons because it has grown up (00:07:02) with that. (00:07:04) That is one lane of security shall we (00:07:08) say. But with the biological issues, (00:07:12) that is far more dangerous to my eye (00:07:14) because they didn't grow up together. (00:07:17) And AI has the capacity to make (00:07:21) biological weapons and chemical weapons (00:07:25) on paper. (00:07:28) Does that make sense? Mhm. (00:07:31) >> Because a lot of AI is pulling from (00:07:35) information in the public domain (00:07:38) and so far no (00:07:42) student in a basement that we know of (00:07:44) has made a nuclear weapon. Think about (00:07:46) that. (00:07:47) >> Mhm. (00:07:47) >> No, it remains this jealously guarded (00:07:53) recipe. (00:07:57) Pakistan got the bomb because they stole (00:08:00) the in most people get the bomb because (00:08:02) they steal it. (00:08:04) But bi biology, we have biological (00:08:08) um synthetic biological situations being (00:08:11) made by (00:08:12) you know students in high school because (00:08:15) of AI because what you can program AI (00:08:18) make me a chemical weapon (00:08:21) >> that to me is a majorly existential (00:08:26) threat. (00:08:29) Hey folks, our friends at Tractor Supply (00:08:31) have always stood for community, for (00:08:34) neighbors helping neighbors, and for the (00:08:36) people who keep life out here moving (00:08:38) forward. And at the heart of every (00:08:39) community are the heroes who serve, the (00:08:42) veterans, active duty military, first (00:08:45) responders, and the families who stand (00:08:48) beside them. These are the folks who (00:08:50) answer the call, who protect, who (00:08:52) sacrifice. And Tractor Supply believes (00:08:54) that that deserves more than just a (00:08:56) thank you. That's why they created the (00:08:59) Hometown Heroes program. It's their way (00:09:01) of giving back in meaningful everyday (00:09:04) ways. As a Hometown Hero, you'll enjoy (00:09:07) special perks through Neighbors Club, (00:09:09) like earning rewards on every purchase, (00:09:12) free trailer rentals and delivery, and (00:09:15) exclusive savings on holidays that honor (00:09:17) your service. And the best part, once (00:09:20) you sign up and verify through ID.mme, (00:09:23) it's automatic. No hoops, no hassles. (00:09:26) So, if you or someone you love is a (00:09:28) hometown hero, visit (00:09:29) tractorupply.com/hometown (00:09:32) heroes to learn more. It's Tractor (00:09:34) Supply way of standing behind those who (00:09:37) always stand for us. (00:09:41) >> But again, we don't have the language (00:09:44) yet, just as layman or with a little bit (00:09:46) of knowledge to understand what AI is (00:09:49) really capable of. And so you're (00:09:51) bringing up the question, should we (00:09:52) trust the same people that said like (00:09:54) duck and cover and you'll be safe? (00:09:56) Should that's what you're saying and (00:09:58) that's a very important question. (00:10:01) >> I would say probably not. We should (00:10:04) probably not trust them. Um I mean I (00:10:07) don't you know I don't know. Once again (00:10:10) this is like the disarming of the nukes. (00:10:12) I mean what do what do we do? We do we I (00:10:15) mean what is your opinion? What do we do (00:10:18) >> about the nukes or about AI? with AI. (00:10:22) >> I mean, I always start by (00:10:25) looking at the opinions (00:10:28) of (00:10:30) people I respect. (00:10:31) >> Mhm. (00:10:32) >> And then I start to kind of (00:10:36) gather more information like why did (00:10:38) they wind up with their opinion? (00:10:40) And so one person that comes to mind (00:10:43) when I was when I was looking at early (00:10:45) AI because a lot of the early AI comes (00:10:47) from DARPA (00:10:50) and (00:10:51) they have DARPA has always had this (00:10:54) idea. (00:10:56) So also I think it's important to make a (00:10:59) distinction at least to my eye between (00:11:03) or I do when I think about it AI (00:11:05) artificial intelligence and machine (00:11:07) learning. (00:11:09) If your machine learning is making (00:11:10) computers a lot smarter, artificial (00:11:13) intelligence is actually (00:11:16) trying to figure out how to make a (00:11:17) machine think. (00:11:19) For that, I visited I went to Los Alamos (00:11:22) when I was reporting the Pentagon's (00:11:24) brain and visited a DARPA scientist who (00:11:27) had a grant to try and create, you know, (00:11:30) a brain in essence. And he was using the (00:11:33) computer that used to have all the (00:11:34) nuclear codes on it. It was really (00:11:36) interesting. But he explained to me his (00:11:38) name was Dr. Garrett Garrett Kenyon and (00:11:41) he gave me this analogy where he said (00:11:43) we're so far out from brain from (00:11:45) computers being able to think and I said (00:11:47) try and just give me a average Jane or (00:11:49) Joe way to understand this and he said (00:11:53) okay think about the this facial (00:11:56) recognition software on your iPhone (00:11:58) right (00:12:00) >> very basic thing you and I that is (00:12:02) machine learning (00:12:04) so he said have your iPhone look at (00:12:09) and then try to have it look at you (00:12:10) further away and with a baseball cap or (00:12:13) with sunglasses, (00:12:16) right? So, you're kind of making it (00:12:18) harder for the machine to know it's you. (00:12:22) Now, this interview we did, by the way, (00:12:24) was like 8 years ago, and things have (00:12:25) changed a lot in a frightening manner. (00:12:28) Then he said to me, (00:12:31) "The iPhone could definitely not (00:12:33) recognize me across a football field (00:12:36) walking with a baseball cap on." (00:12:40) He said, "My daughter, on the other (00:12:42) hand, who is I think something like 8 (00:12:43) years old at the time." He said, "My (00:12:45) daughter knows who I am, across a (00:12:48) football field walking with a baseball (00:12:50) cap, (00:12:51) >> good (00:12:52) >> and begins running toward me. (00:12:55) That is human intelligence. (00:13:00) I mean, do you think I I think I think (00:13:02) we've I mean, look, now China (00:13:06) supposedly has (00:13:08) camera systems. (00:13:11) I don't know what you'd call them. I (00:13:12) guess they wouldn't be facial (00:13:13) recognition. They would just be (00:13:16) >> They are facial recognition. Yes. (00:13:17) >> Well, I guess what I'm saying is it can (00:13:19) pick up how you walk. (00:13:21) >> Gate recognition. And so people are (00:13:23) putting rocks in their shoes so that (00:13:27) they walk different. (00:13:29) >> Wow. (00:13:29) >> Did you know? I (00:13:30) >> did not I did not know the the the way (00:13:33) to spoof that. (00:13:34) >> They're putting rocks in their shoes so (00:13:36) that they walk different so that the the (00:13:38) the (00:13:38) >> the technology doesn't pick up how they (00:13:41) actually walk. (00:13:42) >> Yep. (00:13:44) And they can now (00:13:47) have systems that can read your (00:13:49) heartbeat. (00:13:50) >> I didn't know that. (00:13:52) So this gets into tricky my opinion on (00:13:57) this or rather my lots of facts you know (00:14:00) opinion gets into a tricky area here (00:14:04) because I speak often about the (00:14:06) military-industrial complex not in and I (00:14:09) want to preface this what I'm about to (00:14:10) say about China and all that right which (00:14:12) is not (00:14:14) in a what might be called a (00:14:16) conspiratorial way per se like the (00:14:18) military-industrial complex literally as (00:14:21) a factbased military industrial complex (00:14:28) and it is real and it is also provides a (00:14:32) lot of jobs for a lot of people. (00:14:36) I often think about this. So the (00:14:38) military-industrial complex as a term (00:14:40) comes from Eisenhower's farewell speech. (00:14:42) Okay. And that's very well known. But (00:14:45) less known (00:14:47) is what Eisenhower said as a follow-up (00:14:50) to that in that same speech, which is (00:14:52) interestingly something I you I tell my (00:14:54) sources as a principle I work from. (00:14:57) Would they let me interview them? And (00:14:59) most of them say yes on that on those (00:15:02) grounds which is this that (00:15:05) the way in which (00:15:07) America can function as a sort of (00:15:10) peaceful nation and a democratic nation (00:15:14) and a nation that is strong that has a (00:15:16) strong defense (00:15:18) is through an alert and knowledgeable (00:15:21) citizenry (00:15:23) which is exactly what we've been talking (00:15:25) about this whole time which is in a way (00:15:27) the question for the original listener (00:15:30) asked right how so what Eisenhower was (00:15:33) saying to us is be alert and be (00:15:36) knowledgeable (00:15:38) and so I think it's always good to (00:15:41) temper that like if you say I'm being (00:15:43) alert and I'm being knowledgeable in (00:15:45) sort of like a nerdy way then you can I (00:15:48) can differentiate my like my (00:15:51) pontificating about what does that mean (00:15:53) and I can see my sort of more paranoid (00:15:55) brain thinking thing right and it just (00:15:58) is It it balances things out. (00:16:00) >> Mhm. But on the concept of the (00:16:02) military-industrial complex specifically (00:16:03) and China's surveillance, I want to say (00:16:05) this, which is that (00:16:08) one way of looking at that which I would (00:16:10) look at because I've done quite a bit of (00:16:12) reporting on it is that it's that (00:16:15) problem of the chicken or egg scenario (00:16:17) that that when the United States creates (00:16:19) a radical new technology that it's using (00:16:22) for its own defense, China follows suit, (00:16:26) Russia follows suits, and nowhere is (00:16:28) that more specific. and more obvious if (00:16:32) you really think about it than what the (00:16:35) United States did during the war on (00:16:37) terror, what the government did during (00:16:39) the war on terror and that is create (00:16:42) these these biometric surveillance (00:16:43) systems which you know to go after bad (00:16:45) guys (00:16:47) in Iraq in Afghanistan fingerprint (00:16:49) technology find the bomber not the bomb (00:16:53) a great idea if you you're just going to (00:16:56) take out you if you're going to go after (00:16:58) the the bomb bomb, you're just going to (00:17:00) be think about that's what your teams (00:17:03) were doing. (00:17:04) >> Mhm. (00:17:04) >> But as soon as you can go after the (00:17:05) bomber, you're cutting off the head of (00:17:07) the snake. (00:17:09) But the biometric surveillance system (00:17:11) got out of control before you knew it, (00:17:13) perhaps because of the military (00:17:14) industrial complex. (00:17:17) The Pentagon had decided, well, let's (00:17:19) just get biometrics on everybody. So, it (00:17:22) went from Do you know about this a (00:17:24) little bit? It went I'll keep it short (00:17:26) because it can be like too much of a (00:17:28) rabbit hole, but it went from finding (00:17:30) the fingerprints on the bomber to let's (00:17:31) get fingerprints on every single person (00:17:33) in Iraq. 85% of the population was the (00:17:36) goal. And then they did that in (00:17:37) Afghanistan. These are facts. This is (00:17:40) like David Petraeus fact. (00:17:42) >> Okay. (00:17:44) And so (00:17:46) the idea was we're going to have this (00:17:47) colossal database of everybody, (00:17:51) which used to be considered an FBI (00:17:53) criminal concept. Mhm. (00:17:55) >> We're just going to have this on (00:17:56) everybody and then that way we're going (00:17:57) to know if you're a bad guy or a good (00:17:58) guy and it got totally out of control (00:18:00) and it happened too fast and there was (00:18:02) so much money being made that it just (00:18:04) became a deluge (00:18:06) of systems and China copied that. China (00:18:10) did not have that system of systems (00:18:12) until we introduced it to them. And (00:18:14) because China is great at stealing our (00:18:16) intellectual property, that is precisely (00:18:18) what happened. And then China, because (00:18:22) it's a communist country and it does not (00:18:23) have any of the same rules to abide by, (00:18:26) just went berserk with it and said, (00:18:28) "We're gonna now do they have a system (00:18:30) called physicals for all." Physicals for (00:18:33) all. What a great euphemism. What it (00:18:35) means is we're going to get your DNA. (00:18:38) And that's what they are in the process (00:18:40) of doing. (00:18:42) having DNA, fingerprints, iris scans, (00:18:47) gate monitor of everybody. (00:18:50) So, it's it's becoming a massive police (00:18:53) state if it wasn't already. It is now. (00:18:56) It's a technology based police state. (00:18:58) But remember, my point in that would be (00:19:01) the defense department (00:19:03) set that up, (00:19:05) you could say, to happen. (00:19:11) or is that military-industrial complex? (00:19:15) >> It's a good point I've not thought of. (00:19:18) >> It's a great point. (00:19:21) What do you think? (00:19:24) >> I think it's the Eisenhower quote like (00:19:26) an alert and knowledgeable citizenry, (00:19:29) right? And and also a little bit if you (00:19:31) pick your battles because you could you (00:19:34) can become subsumed with (00:19:36) this is just a horrible you know you (00:19:39) could really um and you want to enjoy (00:19:42) your life and be a good parent and write (00:19:44) your books or do your podcast. (00:19:46) So and then I look to history to say (00:19:49) okay oh that's right this has always (00:19:50) been going on. (00:19:54) I do believe money money needs to be (00:19:57) spent to keep the economy going, but (00:20:01) there could be a restructuring of the (00:20:03) military-industrial complex (00:20:08) in a manner that (00:20:10) suits (00:20:12) the livelihood (00:20:15) and the future. (00:20:16) >> Yeah, I agree with you on that. No (00:20:19) matter where you're watching Shawn Ryan (00:20:21) Show from, if you get anything out of (00:20:22) this, please like, comment, subscribe, (00:20:26) and most importantly, share this (00:20:28) everywhere you possibly can. And if (00:20:31) you're feeling extra generous, please (00:20:33) leave us a review on Apple and Spotify (00:20:36) podcasts.