What Does Artificial Intelligence Mean?

00:00:07
Speaker 1: Alexa, what's the best science podcast on air? Hey? Are you trying to replace me with Alexa? What's going on here? Do you think you're replaceable? There's no way an artificial intelligence could ever make jokes nearly as funny as I am. I think there's no way an artificial intelligence would laugh at your jokes. I'm pretty sure I could. I could program a pretty dumb computer to laugh at my jokes. It's called the laugh track. But that's Hey, that's a new challenge for for AI. You know, first chest then go now science comedy. That's right, now, programs something that can find humor in Daniel's ramblings him and I'm Daniel, and this is our podcast. Daniel and Jorge explained the Universe, in which we try to download everything we know about the universe, episode by episode into your brain, whether you're a real person or an artificial intelligence, listening to our podcast while trying to sound intelligent about it while writing your own humor for the open mic AI Night. The topic of today's podcast is what is artificial intelligence? And very importantly is it dangerous? That's right? Should you be looking at your window for the first signs of the robot revolution. Should you be afraid of your Alexa? Should you be worried about that robot vacuum cleaner getting resentful for having to do all the dirty work and eating your face off in the middle of the night. That's a bit dark. It seems kind of sinister, doesn't it. It's like sitting there are circling, circling, circling, waning, waiting, waiting. I think those things are creepy, right. Maybe it wants to, you know, clean your face, it wants see That's the question. Does a robot vacuum cleaner want anything? What does it mean for it to want? What is it like to be a robot vacuum cleaner? The next great paper in philosophy? So this is kind of in the zeitgeist right now. I mean, people are really excited about artificial intelligence. But at the same time there are big names like Elon Musk kind of warning people like, hey, artificial intelligence not such a good idea. That's right, it's a huge topic. I mean, you drive around like San Francisco, you see artificial intelligence, machine learning, deep learning. It's on billboards. Even you know you want to get a million bucks for your new company, you just say the words AI, deep learning, and boom people are throwing cash to right, people are learning in the deepest learning. Um. It's definitely part of the cultural moment, and you see that reflected not just in like what deep thinkers are saying, but also in like science fiction. You know, a lot of the near term dystopian these days is about how AI will take over and the dangers of AI. Another way, like thirty years ago is about the dangers of UH of radiation. Right, that was the new dangerous thing physicists that invented. Now the new dangerous technology that we're all worried about is AI. It's a new promise in peril. Um. Yeah. AI. Every piece of technology is a double edged sword, right. You can use it for good, you can use for evil. But AI is special because it's not just technology, is not just a tool that people use. It's a tool that that has independence, that has autonomy. And that's why it's such a vexing question. Well people, I'm sure people everyone associated with robots and machines and computers, but we were kind of wondering if people actually knew what artificial intelligence was, like, what makes it work, what makes it different than than real intelligence? M I bet you that the people who say the phrase artificial intelligence don't actually know what they're talking about, which is probably true for most topics andology, it's true for me probably, I'm sure. But we're wondering if you guys out there knew what artificial intelligence was. And so, as usual, Daniel went out and as people in the street, and here's what they had to say. Um, it's the idea that we can create some type of material thing that could think on its own ultimately, And do you think it's something we should be concerned about? Is it ever going to be a threat to humanity? I mean possibly, but I mean we never We don't know everything. We can know the bounds of what could be a threat, we cannot be a threat. Yeah, it's AI, and it's the stuff that's used in various technological applications basically just kind of like trying to make machines replicate certain aspects of human intelligence. Stuff like that. Okay, And do you think it could ever be a threat to humanity? Is something we should be worried about? I guess since I don't have a particularly strong opinion on it, I don't think so. So I guess I'll say no for now. Um, I'm assuming that's the idea that computers or electronics can have like sentience. Right, Are you worried that computers would when they take over and make us their slaves? And not really, I don't think it will come to that point. All right. Those are pretty sophisticated answers. I like the ones that said, um oh, artificial intelligence, that's just AI, right, Like that's an answer. So that's an answer to every question. You know, what is Google blex Zavi Brown? Oh, that's just gens. Yeah, acronyms. Acronyms can make you look intelligent, that's it. That's the real artificial intelligence to speaking acronym acronym intelligence. Um. You know, but people had some sense that it's, you know, something that can think for itself, or something do something for you, or create something that can think by itself. There's definitely the nuggative idea is definitely out there. They use it in relation to what it can do. That's right, yeah, exactly what what is what's the new capability that defines it? Yeah? Right yeah, and that it's it's a fascinating way to think about it, you know. And uh, it's definitely a tricky question, right because I guess we know it in the context of UM using them for things, right, Like, we're people, don't just create a I because we want to create artificial beings. It's like, so it can help us. I want to create artificial beings. What's what's wrong with that? That sounds pretty awesome. Create a whole army of physics artificial physics grad students. It sounds pretty cool. Kids. Yeah, you mean, are they worried about competing with my digital children? Natural? They know you'd rather have artificial children. I didn't say I'd rather have artificial that in addition to my beautiful, wonderful natural children, which I should not be talking about on this podcast. I'd love to have a whole, you know, cadre of artificial children to do my bidding, unlike your real children. If who won't do your bidding? So somebody listening children? And that sort of goes to the heart of the question. You know, UM, if you created a digital being with artificial intelligence, would it listen to you? Or would it make its own decisions? Right? And so that's why we thought it would be interesting to dig into, like what is artificial intelligence? If it just did what you told that to do. It wouldn't maybe be in an artificial intelligence you're saying nobody smart should listen to you, is what you're saying. I'm saying they should decide for themselves whether I'm I'm worth following. So let's break it down for people. Daniel, what is artificial intelligence? Well you should listen to this podcast and that will give you the answer. Done. Um, Well, you know, I think to understand what artificial intelligence is, we should think for a moment about what do we mean by intelligence? Right? And very simply intelligence. It's just the ability to learn, is to find patterns to extrapolate from them. Really, that's how you. But like a dog can learn. But a dog, you wouldn't say it's intelligent, would you? Absolutely? I would say a dog is intelligent. You can teach a dog, you can train a dog. It's more intelligent than a rock. But would you say by a lot, Oh my gosh, if you like, never interact with a dog. A dog is like a living sension being. It feels that experiences, It definitely learns. It can recognize you. I mean, dogs can do complicated of think. The dog is a perfect example it's you know, I wouldn't trust it to do my taxes. You know, I don't know compared to our tax accountant, did I do a pretty good job. I mean, you can say that's an intelligent dog, but you wouldn't say, like, that's the epitome of intelligence. I wouldn't say the dogs are the most intelligent beings in the universe. But that's what we're talking about. We're talking about do they have intelligence? Pretty example, because they can learn, you can train them, and you the cool thing about an intelligent being is that you can train it to do something even if you don't know how to do it. Say, for example, you want your dog to recognize you, right, but tear the face off anybody who tries to break into the house, right guard dog. Okay, so you can train a dog. You reward it when it does the right thing, and you punish it when it does the wrong thing. You don't know how to, like build a being that does that, that like recognizes your face and recognizes stranger's faces and makes these decisions. That's hard task, you know, it's not easy to do. But you can train a dog. A dog can learn how to solve this problem and all you need to do to train it is to reward it and punch it. So you're saying, just the ability to sort of learn from your mistakes or learn from your surroundings, that's what you would call intelligence. Yeah, And dogs have less of it than we do, and more of it than cats and mice, um. But they have some of it for sure, which is what makes them trainable. And you know, I wonder sometimes, because dogs can be trained, right, nobody ever trains their cat. What does that say about a cat's intelligence. I've always thought I love cats, but I've always thought dogs are probably smarter than cats because you can train them, right, Or maybe cats are more intelligent in that they're they're not they don't allow themselves to be trained by humans, right. And rocks, by that metric, are the most intelligent because they completely ignore you, right to see the fallacy of that argument right there. But I mean, maybe there's sort of sort of like a hump, right, like, as you get more intelligent, you're easily more trainable, trainable, trainable bout somebody. You get so intelligent that you rebel against your masters, And so how do you tell the difference between something that's totally unintelligent and something that's so intelligent and completely ignores you. Yeah, I don't know deep question or believes all the rocks are probably thinking about him. Well sure, I mean if he used to use the ability to listen to what I say as a benchpark of intelligence and then yeah, there's, um, something super intelligent could be just as smart as the rock. But obviously a cat is still making decisions and acting and you know, doing things, so it's intelligent. But maybe it's much more intelligent than a dog because it's it chooses not to listen to us. All right, I think we need to have a whole other podcast on who's smarter cats or dogs? And before we do that, we will collect some data to answer this question. Um, but I think with the question we're focusing on is what is artificial intelligence? So natural intelligence just the ability of an animal to learn. Artificial intelligence would be if something artificial that we create has that same property, the ability to change the way it processes things in response to what it sees about the world. Yeah, artificial intelligence is a very broad field with lots of elements that we couldn't cover in just one episode of a podcast. But let's just talk today about one important sub field of AI, which is machine learning, or more specifically, I would say, let's focus on training. Right, can you build something or something artificial that can be trained? Right? And uh? I think, I think let's talk for a moment about, you know, how how normal computers work, and then we can talk about how computers, smart computers, computers that can learn, computers with artificial intelligence, how they work. I think, I think what you keep talking about cats and dogs? All right, we'll talk about cats and dogs, but first let's take a quick break. Let's talk about what computers can do. Yeah, let's because computers are smart, right, you can program a computer to do smart things, but that doesn't necessarily mean it has intelligence. That's right. There's a difference between a computer that can do something and a computer that can learn something. Right. The way I think about non intelligent computers is the way you sort of think about machines. Right. You can tell them what to do, and they do exactly what you tell them, regardless of whether it's the right thing. You don't give them like a goal and say, hey, I just want the house to be clean. Figure it out. You have to tell them exactly what to do. You say, step over here, move the broom this way, step over there, you know, and if it's not cleaning the house because they're stuck on a corner or they're you know, fell on their on their butts or whatever, they don't care. They just tell you do exactly what you tell them to do. Have no sort of larger sense of what's important. It just follows instructions, just follows the recipe you gave it. That's right. It's like a like a wind up toy, you know, you wind it up, you give it some energy, and then it goes. And I really do think about computer programs the way you might think about little machines, right, because that's exactly what they are. They just execute a set of instructions. You know. It's just like a bunch of gears clicking into place, and they can't change the way they do that. And they do it regardless of whether it's the right thing, or whether it's effective or whatever. It just goes. Like your electric toothbrush, you know, you switch it on, and it's just it has a circuit that just has it moved the bristles back and forth, that's right, And it doesn't know if it's brushing your teeth or just flailing around in midair. Right, has no idea, It doesn't care, It doesn't think or feel whatever. It's just a machine, right, Thank god, it doesn't know. It would tell you to brush your teething like that's chocolate, Jorney, I'm tired of this? What is this gunk? Yeah? Exactly. And so that's what a sort of a normal machine is. That's what a classical computer program is, right. I think of it just the same way as you think of a physical machine. Okay, it's just doing what you the programmer told it to do. That's right, and it follows your instructions exactly. Um. Now, a computer that can learn is different, right. A computer that has artificial intelligence is different in this really important way because you can train it right, and you can train it because you we build these things to model the way that we work. Right. So for example, an AI program is sort of like, um, like a newborn baby can't do anything. Right. Say, there's a AI program, for example, that's supposed to recognize you when you come in the door. Right, is this Jorge or is this not Jorge? Right? Because it should only open the door for Jorge and not open the door for not Jorge. Okay, So when you created a new AI program, you would start out like just a newborn baby, okay, and like a blank slate, right, Like a blank slate, it would make random decisions, right you You you show it a face and it would say yes, it's Jorge, and then you say no, you were wrong or yes you were right. And then you would reward it if it does well, if it gives the right answer, and you would um, you would punish it if it doesn't, right, you would tell it. I mean, you don't actually punish it a reward. You just tell it, yes, you made the right um call this time, and know you made the wrong call this time this other time. But how is that different than the of calibrating something? Do you know what I mean? Like is calibration than artificial intelligence? Right? Well, the difference is calibration is like here, I have a tool. I know how to solve the problem. I just have to adjust it so that it does exactly the right thing. Um here, right, But you have a strategy that it's executing. You know. It's like, uh, you have a drill and you want it to drill fast or slow, and you know you know what how to solve the problem. You just you know it has to spin and screw and screw the thing in or whatever. It's just adjusting a knob here. You don't know how to solve the problem, and so you've given it a very very very flexible strategy on on the inside. You've given it like imagine something has like a thousand knobs. If you twist all these knobs, you could get all sorts of crazy um strategies. Right. So back to the example of like recognizing Joge or not when it when you tell it it's done, it's given the wrong answer, then it it adjusts those knobs. It says, well, let me try to tweak my strategy for deciding is this Jorge, and then we'll see how that goes. I think that's a key difference. It's the number of knobs, right, Like a drill with the knob for velocity. I mean, that is sort of trainable and you can set it up to be adaptive. But it's just one knob, and so it's not You wouldn't say it's intelligent. It's not intelligent. The spectrum of things it can do is very very seene right. But whereas like something that recognizes a phase. It needs to evaluate like a million pixels in a photo, right, and so for you to tweak how it evaluates each of those pixels, it would be really difficult for you to That's right. So imagine you know the machine here is is a camera in the door and takes a picture of who you got a million pixels and then it has to look at those pixels and decide is this Jorge or is this not whohe? And so there's does some calculation on that picture, right, and that calculation has millions of knobs on it, right, how much do I weigh this pixel? How much I weigh adjacent pixels? Do I look for his nose? Do I look for his hair? Do I look for the eyes? Right? So it's got some very very flexible thing inside of it that can do almost anything. And when you first start out, it's just random. So it's making ridiculous, terrible decisions. But the key the thing that models the learning, right, you know, just need artificial intelligence, You need artificial learning. The thing that models that learning is that when it gets the wrong answer, it knows how to adjust those knobs so that next time it's more correct. By itself. That's the key thing is that it learns by itself. It doesn't need you. They're sitting like, oh, you got this pixel wrong. You've got that pixel wrong. A tweak you know, tweak this one this way. It's really more like an autonomous automatic learning. That's right because you don't know how to adjust it. If you knew how to adjust it, you would just write that program. Right. The key is artificial intelligence is excellent when you don't know how to solve the problem, but you can define the problem. You can say, this is a picture of or and this is not learn a way to tell the difference, right. So you give it a very flexible strategy and then you try You let it try out, and when it gives it the wrong answer, you would let it adjust itself so that it gets closer and closer to giving the right answer right, And eventually these things will find the right setting for those millions of knobs, so that it's doing the right thing. It's saying, oh, look, this picture is a picture of Horhey, and it gets the right answer at the time. And when you give it a picture that's not a picture of Horne, and you give it Daniel, it says no, sorry, you're not getting in the house. Right. And I think a key thing is also that you, as a programmer, could not have predicted what all those knobs are going to be at the end, right, Like, it's such a big problem. There's a million knobs. There's no way that you can predict what those knots are going to be set to when it learns my face. That's right. It's perfect for really hard problems where we don't know how to solve it, right. We know how to describe the problem, but we don't know how to solve it. You're right, So if I already knew how to solve it, I could write a computer program that that and and tell it just like use this pixel, use that pixel, use this pixel. But I don't know how to solve that problem. It's really hard, right, But I can train a computer to figure it out, just the same way I can train a dog. Right. A dog can learn my face. Right, a dog recognizes the owner, and you know, happily licks its face when it comes home, and recognize that that when somebody's not its owner, and barks like crazy and choose its face off when it's not its owner, Right, remind me not to visit your house, Daniel, seems a little dangerous. So then a big thing is programming a structure in the in the software that is kind of open ended and malleable, do you know what I mean? Like, its something that is kind of unpredictable in a way that can learn. That's right. And that's the key thing is that some people might be thinking, well, hold on, you said that the computers can just do what they tell you, So how can a computer learn? Right? How is that possible? And the key is that it's an emergent property, right. Like the way that you write a computer program that can learn is you build all these little um calculating bits with knobs on them, right, and each bit just does what it's told. It takes some data, it makes a decision based on the value of the knob, and it sends out some data. And together all these things may a decision. Right. Each individual piece has no idea what it's doing. It's not smart or intelligent or making its own decisions. That doesn't have free will, right, But together they're doing something. And as you said earlier, they can change the way they behave. They can adjust these knobs themselves to improve their performance, and that's where the learning comes from. It's from that training. It gets external input and changes its behavior based on that expe. You're saying that the way to program these AI s is by h connecting a bunch of little simple things together to get something complex. Yes, And here it's important to remember that we are using neural networks as sort of a standing to represent a big broad set of strategies that are part of machine learning, right, And you don't know how to set them, how to put them together to get the right complex behavior. You just put them together and then you train it. Right, You say, well, I have I have something that's dumb, like a newborn baby, and I teach it how to do the thing that I want. But this really all sort of came about from brain research, right, Like, people were studying the brain and they figured out that our brain vans are made up of all these little simple units neurons, that's right. And each neuron is pretty simple, right, Like, it just takes a couple inputs and then it just outputs one signal. That's the really fascinating deep part about it, right, is that the structures we use in computers are modeled after what's actually happening in real brains. And when you say, inside your brain are a bunch of neurons, right, and these neurons taken some input and then if the input is right or above some a certain amount, then they send us some output, which is the input to the next neuron. Right, and your brain is basically just a big web of these things. Yeah, yeah, that's the right. That's the key is that these neurons, they're simple, but they're all sort of connected to each other. So it's a huge complex web going on inside your head. And when you're learning, what you're doing is you're kind of like shaping that web. You're saying, some connections. These connections are important for recognizing kore. Uh, these connections are important when you want to when it's not where kind of think. You know, your neurons can change. They have like basically knobs on them. I mean not physical literal knobs, but they have they can adjust. And so if you feel pain, you know, or you have an experience, then that changes the way your neurons work and it changes a little bit who you are and how you react to things. And that's why you know, newborn babies when they're born, they're not very responsive to stimulan because they're just still figuring figuring it out. You know, a newborn baby doesn't even know like, this is my arm, and I know how to control It has to learn all of these things by being trained, by having experiences. You know, it has the neurons, and the neurons are connected to each other, but it hasn't figure out how to use those connections. I try. It has to be trained to be useful and to interact with the world in any sort of meaningful way. Right, And so that's exactly the same sense. And it's fascinating that if you build a mathematical system. That's what a computer program is, basically a mathematical model of the processes that are happening in your brain. It performs in a very similar way, and it does this amazing thing, which is it adjusts itself to improve its performance on the task you've given it. Right, So it really is like a model of learning. And and when people saw this, they said, wow, I mean you you look inside the brain. You're wondering, like, how does thinking work? Where's the soul? Right? Where am I? You look inside the brain? All you see all these weird neurons connected to each other. You think, how could that possibly describe me? But when you build a model in a computer and it can do the things that you can do, which is you learn and develop and react and be trained and make bad jokes. Not yet, we have not yet solved the bad joke problem. Right, humans are still world champions in terms of bad jokes. We can still beat them at something that's right. And you know, and this is very useful because you want the systems around you to learn and to react, you know. And like if your phone, for example, it knows, hey, every time you open your phone, you start with Twitter, right, and so Twitter goes up on there on the most used app list, right. And that's not very complex artificial intelligence, but it is. And uh, and these sort of things are very helpful. Let's take a break. So that's kind of what makes AI is that a it can tackle complex problems that we don't wouldn't even know how to program something to do, and be that it changes and adapts and kind of it can get better, not just better, but also kind of adapt to the person using it. That's exactly right, exactly right. And so for example, sometimes you know Netflix uses AIS what program will you want to watch next? Well, you know, um, that's an AI. It's been trained. They feeded a bunch of examples. They say, Bob watched these five shows and then he watched this sixth show. But they gave the AI just the first five and they asked it predict what show he will watch next, and then they see if it doesn't do a a good job, and if it does a good job, you know, they reward. If it does a bad job, its its knobs to do better. Then when you're sitting there watching five hours of Netflix, it can do a pretty good job of predicting what you're gonna watch next because it's been trained on a lot of data. This is why people always talking about big data. Big data. These companies are gathering data about you so they can train their aiyes to learn your behavior and predict it. Except the problem is me and my spouse we share the same account and the same log in, So that's right, So it's learning some weird in your wife's brain. I have a very confused Netflix. I can or maybe it understands your marriage better than maybe's trying to tell us something. It's like you guys. Should you guys? His wife is out of town, he watches these shows. When she's in town, he has to watch these other shows. Oh alright, I know, break it down for us. How long before the aiyes take over the world? Um? Not very long actually, But you ask me different question earlier, which is is AI dangerous? And I think that has the two different questions there, right, I mean people are concerned. Some people are concerned. Yeah, I think people are concerned, and they're a good reason for it to be concerned. You know, One question is will AI develop its own autonomy and uh, and you know take over. That's a different question from are they dangerous, because you know they could take over and then take better care of the planet than we have, in which case you know they're not dangerous. They're benevolent dictators. I think the real question is will they take over? Will they become autonomous? We lose control of them somehow? And could they become smarter than us? So I see it's two issues. One could it could have developed a consciousness on its own and be is that consciousness good or bad for us? And it's an important question because as soon as you identify learning with consciousness, right, then you wonder about that and and this connection between the structure of AI and the structure our brain begs that question. You know, if you created, for example, and artificial or hate in the computer, if you built a set of neurons that mimic your brain, you know, would that simulation be alive? Would it be aware? Would it think I would have a first person experience? You know, that's a deep philosophical question, will never answer, right, And it's not really the important question. The important question is would we lose control of AI? Well? AI, because AI is something that can change and that can evolve, It can handle complex tasks. The question is can we lose control of it? And I think the answer to that one is definitely yes. We can lose control, meaning like, um, we'll give it control and then not be able to take it back. Yes, exactly, because the way AI is moving is that it can handle more and more complex tasks so that you don't have to be super specific about what it's doing, you know, like we have amazing natural language processing. Now you can say sort of vague things to your phone like hey, UM, set me up an appointment for tomorrow afternoon, right, and it will understand because it understands what your intent was, right, has to judge your intent and then execute it. It used to be you have to go into your computer and you have to press the keys in order to create that in your calendar. Now you can sort of talk to your phone and it will interpret what what you want and it will do that. And that's that's awesome. That's wonderful for human computer interactions that we can use our language to talk to them. We don't have to write computer code. That's a huge step forward, right, that people can construct machines using English rather than Python or C plus plus. Right, it's a big step forward. I think that's kind of what people find scary about a is that you can't really predict what it's going to do. I mean, it's sort of comedy gold when your kids are trying to talk to Alexa and ask it funny questions. But that's kind of what's fascinating about it, right, Like you ask it questions, do you give a task and you're so you really sort of don't know what it's going to do. That's exactly right, because it's making higher and higher level decisions, which makes it much more useful and much more intelligent. The same way when your kid grows up, right, when it's a When your kid is for you have to be very specific. You have to say things out loud which are ridiculous us, right, like don't put that finger in your nose. You know, we're like, uh oh, that's been on the floor, don't eat it. Right, you have to be really specific. When they're ten, you can say more general things and they'll understand, right, and learned their intelligence, like don't put both fingers in your nose. Only put one finger in your nose at a time, so you don't put your finger and your sisters and right, um, the same way as machines or artificial intelligence gets more intelligent, you can give it vagrant instructions and then it makes decisions based on its training. Right, and you don't really know, just like you can't really read what is in every neuron in another person. And then AI you sort of you don't know what's gonna happen, what's gonna come out exactly, So they're gonna start making decisions based on you know, still what we tell them to do. But you know what if you told your AI, you're like, hey, keep my kids safe right. I mean, imagine some future where you have an AI robot it's really smart, and you say, hey, keep my kids safe, and you come home and it's like lock them in the basement, right, and like, well, okay, they're safe. But it's sort of a monkey pause situation, right, Like you got exactly what you asked for, but you didn't really labor the right way and made different decisions. Right, So we still skip the question of whether aies can be you know, a chief consciousness and become its own kind of soul, have a soul. It doesn't seem like you think that's a relevant question. I think it's important because when AI gets to be super intelligent, it's gonna seem like it has a soul. They're gonna seem like people, and people don't wonder like do they have rights? Can you kill an AI? What? How can you just delete it? You know? Um, that's going to be a really interesting question. But that's again that's a whole question of philosophy that we could we could easily spend an hour on things a much more practical question, which is will we lose control of them whether or not they have first person experiences so they just seem to. It's important to think about whether we're gonna lose control. And there's two reasons why I think that we will. One is, computers are getting faster, really really quickly. Right, Every year, computers get faster and faster and smarter and smarter, and the scale is is is growing, right, So this thing is happening very quickly. But we're not right. We're not getting smarter. Right, human brain is not changing and evolving at a very rapid rate. Computers are, So they're catching up and the slope is steep. Right, you can just get bigger and bigger computers and teaming together and paralyze them and you can just keep going right, So eventually they'll definitely have enormous computing power with capabilities to do things we can't even imagine. And also being faster doesn't necessarily mean being smarter. You also need like more dated to train on it. And also being twice as fast doesn't mean being twice as smart. It's not linear. So you think that they will get more capable than us. But do you think we will ever seed control of really important things to Aiyes, Like, hey, here's the nuclear button, um only fire it if it's necessary exactly, Let's talk about weapons. Weapons is going to be what ends it because you know, for example, we already have drones, right, and we have drones with missiles on them, and these drones can kill people. They can, they can, you can. Some pilot somewhere is flying it. He's making a decision and he's gonna shoot dismissile to kill a person, right, right, But you know the enemy has drones, and pretty soon it's gonna be drone on drone warfare, right, And again, drones are gonna shoot each other, and at some point somebody's going to put an AI in their drone. Why because an AI can make the decision about shooting much faster than a human can, So which drone is gonna win? An AI will be a better fighter than a human fire yes, And so eventually these AI will be making kill decisions, right because the one that can make the decision faster, it's going to be the one that wins. And so I don't think it's gonna be very long before we have AI powered drones that are authorized to kill people. Right. This is a clear next step for the military. You know, like here, here's a picture of somebody we you think is a terrorist. If you spot them, just fire the missile. Don't bother checking, Yeah, don't bother checking with us. Right, that's a clear next step. So now you have AI that have the authority to kill people, and why because they've been tasked to, you know, take care of us or protect us or only only if you give it that permission though, right, Like, I mean that's a big ethical step to say, like, if you see him, shoot him. Yeah, But I don't think that's a big ethical step for the military. You know, the protocols for shooting somebody in the military. I mean, I'm not an expert on military protocols, but you know, our military kills a lot of people for you know, a lot of civilians get killed, right, and we decide it's okay. A lot of innocent people get killed for military purposes. And so I don't think it's too far before AI is making that decision. And then it's AI. It's weaponized AI, our weaponized AI versus their weaponized AI. And then it's an arms race, and then the most powerful army is gonna be the one that just makes it all of his decisions, and the generals just say defend us right, or respond if we're attacked, right, And then you basically handed over control of the weapons to the AI because the enemy has weaponized AI. But that doesn't mean that they're trilling us. I mean, we use them to protect us or to take away some decision making for it, but that doesn't mean that they're necessarily in control of us. And let's make sure not to be too alarmist here, of course, because people are working really hard to make sure that there are always ways for humans to override these systems. We would be different. That would be um, you know, it'd be like if a robot then turns the weapons inwards. That's another deal, I guess. Yeah. And of course AI researchers do their best to make sure that the AI systems are very well trained so that they do exactly what we want them to do. But they are complex and unpredictable, just like people are. Right, So this is a very interesting topic whether AI is dangerous or not. And I know Daniel that you you're sort of an expert in artificial intelligence because you use it in your particle physics research, right, you use machine learning, that's right. I wouldn't say I'm an expert I mean, I know something about it. Um. I've done some reading and I've used it, but I'm certainly not a deep expert in artificial intelligence itself, right, But you you know experts in your department, right, and you're in your campus, that's right. You see, I has an amazing computer science department and experts in machine learning. Some of the folks I actually collaborate with. When we're understanding the huge amounts of data from the large Hagon collider, we train machines to sift through that data and like look for the Higgs boson and learn to recognize new kinds of particles. It's really fun. And these guys know a lot about artificial intelligence more than I do. So I went over there and I asked them if they were worried about whether robots would take over the world, and what did the robots say the robots had taken over the professors and they answered for no. Um. First, here's professor Pierre Baldi, he's a distinguished professor on campus. And here's what he had to say. Potentially, yes, all very powerful technologies I think can pose such a threat, and all depends how they are deployed, how they are used. Et cetera. Right, you can say that nuclear technology pulls is such a threat and continues to pull such a threat. And I think AI, if used in the wrong way to pose a threat to mankind. Yes, the potential is there and so we should be careful. Um, right, So that was Professor Baldy, and then I also went down the hall and asked another colleagues cause I thought let's get more than one opinion, and so this is Professor Park Smith, also a professor of computer science at U c Irvine. I think the main threat with artificial intelligence going forward is not understanding how the black boxes work. And so I think not the typical sort of we're going to have robots taking over the world, but more the use of AI and situations where we're extrapolating beyond what it can do. And so I think we need to understand the limits of a I I think that's a threat, all right, So the answer is yes, Well, I think they're cautious, right, both of them think it's unpredictable. We don't know what's going to happen. We're creating a whole new kind of system and uh, and we may lose control of parts of it. On the other hand, you know it's likely for that to happen. You know, a lot of people are working really hard to make sure that AI will be contained and that in the end you can just pull the plug if the robot revolution starts, and so it is unpredictable. But also you know, the future is unpredictable, is always going to be unpredictable. Yeah, I feel like I thought it was interesting he said it is dangerous, but not more so than any other powerful technology. Yeah, that's a really interesting comment. It's true that any technology you can create could be used for good or for even if it's powerful, like I mean, not just like you know, wind up toy. Maybe it's not as dangerous. But but but I think that speaks to the kind of the power of AI, Like it really is maybe more powerful than we can handle. Yeah, and it's it's powerful in a special way run like nuclear weapons are powerful. Right, But in the end, a human is making that decision, and so you're giving humans a new kind of power, which is unpredictable. But here you're you're unleashing something, right, You're creating AI, and it's making its own decisions. Of course, it's making decisions based on what has been told to do. Right, you have to give it instructions. Still, you have to teach it um. But you can't predict what these complex systems are going to do in new circumstances and how they're gonna interpret your instructions. And of course there are a lot of AI supark people out there working hard to make sure that their boundaries and safeties um installed in all AI systems. But you know, I've seen Jurassic Park, you know, the lesson there they had fences. Lesson there, they had fences. We have fences. But then Jeff Goldbloom, you know, has a theory about chaos. Yeah, exactly. You know, these systems are hard to predict, and so I think we should be worried, but then we should respond to that worry with appropriate safeguards. You know, we should take this seriously, but not be overly alarmed. Right. Well, the other point that the other professor made is also interesting that he's saying some of the danger is in the fact that it's kind of like a black box, like we're trusting these things, but we don't really know what's going on inside. Like it's so complex they we we can't predict what it's gonna do. We can't maybe even deconstruct how it makes decisions. That's right, and uh, you know you train these systems are very complicated and you don't know how they're gonna respond to new circumstances. Right. It's same as when like training your dog, Like do you know how your dog makes a decision about who to bark end who not to. You try to train it, You try to give it instructions to try to make sure it knows how to how to handle it stuff in a new circumstances, but you can't honestly know what it's going to do at any given moment. Yeah, I'm definitely not visiting your house if you have dogs. I think about I think about AI. Again, not an expert, so maybe these are uninformed speculations, but I think about AI sort of like digital children. You know, like you raise your children, you know they're gonna take over one day because you know, you and I are going to get old and our kids are younger than we are, so eventually they will take over and you don't know what they're gonna do, and you raise them. You try to raise them in a way that they have values they make reasonable decisions, and you can sort of think about AI to same in way like you try to create this new generation of technology that's going to make its own decisions, but you try to teach it to make good decisions so that when you're in a home, right, it's making good choices for you. And I know that some folks out there think, well, you know, AI is never really going to be separate from humanity. There's not this like cognitive separation, Like you can just be part of who you are, the way your iPhone feels like part of who you are. Um, But we don't know necessarily if if that separation is going to be serious, you know, if these things really would be separate from us, or if they always just feel like an extension of ourselves. Well, until then, I think we should stick to regular dogs. Dogs. Yeah, But you know, I think about it sometimes the way I think about children, right, In the same way that you raise your children and they're gonna take over, right, It's gonna be some point when your children are in charge. You raise them to have values and to make good decisions, and you hope that when they take over, they're you know, looking after you. In the same way, we got to create these digital tools, and we've got to teach them to be Hey, we got to teach them what's important, and we got to teach them how to be responsible so that if they take over, you know that we hope they treat as well. Yeah, daddy, good daddy. Your parents don't put creator good Please don't bury me underground. Well, I personally am looking forward to a time when I have, like I don't have to think as much, where life is a little bit easier because we have these things making things easier for us. It could handle a lot of the drudgery and a lot of the logistics. You know, eventually you could have a car that drives itself and obeys your instructions. You can say like, hey, go pick up my kids from school, and he would know how to navigate and how to drive and recognize your children and how to get back home. And that's totally within the realm of possibility in a few years, right, And that's pretty awesome. It will offload a lot of work and logistics from beleaguered parents. I think you and I are in a pretty good position career wise, you know, Like I'm a cartoonist in your physicist. These are not um jobs that are going to be taken away by AI anytime soon. Hopefully have you not seen a our cartoons? They're pretty good man, all right, they you should like start a podcast instead of wearing of relying on your cartooning. Well, there is definitely that as a genre of humor. Like, hey, I put um so and so through an AI machine and look look at the crazy thing it came out with. Except those are all manufactured. None of those are no, those are real, None of those are real. Those are all made up. Well, that's good for humorist. So artificial intelligence is certainly a revolution in thinking and in computing, and it will definitely change the world. And so check back in in ten years to see if we've been replaced by robot Daniel and robot Warhead. Maybe we already are bump bump ball. So thanks everyone for listening to this episode of Daniel and Jorge Explain the Universe, and to listen tomorrow. Just say, Alexa, what's the best science podcast in the world. What's the third best? It's not a Catherine the world. If you still have a question after listening to all these explanations, please drop us a line we'd love to hear from you. You can find us at Facebook, Twitter, and Instagram at Daniel and Jorge That's One Word, or email us at Feedback at Daniel and Jorge dot com