The Daily Show: Ears Edition - TDS Time Machine | Artificial Intelligence

Episode Date: June 8, 2025

*Generate a description for a podcast from The Daily Show about AI, including the following segments: Jon Stewart on the false promises of AI, Trevor Noah on whether AI is self aware, Wanda Sykes on AI Hitler, Kal Penn on the release of ChatGPT-4, John Leguizamo on the warning signs of AI, Ronny Chieng and Michael Kosta on AI being used to unlock ancient history, Dulcé Sloan and Ronny Chieng discussing the rise of AI girlfriends and Trevor Noah interviewing CTO of ChatGPT Mira Murati. Make sure it reads naturally, as if it was written by a human.See omnystudio.com/listener for privacy information.

This is an eye heart. You're listening to comedy Central. There is technology out there in the world that really does blur the line between reality and tailgate art. But those are mostly I generated your fake Joe Biden robo call that tells New Hampshire voters not to vote, your Chicago mayoral candidate

glorifying police brutality, your Donald Trump dropping by the neighborhood for a stupac. Look how comfortable he seems. And as AI gets better and better, it's only going to make it more difficult to separate fact from fiction, which could be terrifying. Luckily, the people in charge of AI have told us that just like with the internet and social media, it's actually going to make everything much, much better. This has the potential to make life much better.

I think it's honestly a layup. I hate to sound like utopic tech bro here, but the increase in quality of life that AI can deliver is extraordinary. AI is the most profound technology humanity is working on, more profound than fire or electricity. Yeah!

Suck a d***,! That's right, you heard me. You heard me, fire. Oh, I'm sorry. Do I need to turn that up? Suck a motherf***ing d*** fire! And oh, whoa. What are you giggling at electricity I Mean listen, I'm sure AI is good but like fire good How so They can help us solve very hard scientific problems that humans are not capable of solving themselves

solve very hard scientific problems that humans are not capable of solving themselves. Addressing climate change will not be particularly difficult for a system like that. The potential for AI to help scientists cure, prevent, and manage all diseases in this century. I completely trust you. And your enormously wide eyes and very human cadence. But benefit of the doubt, this can cure diseases and solve climate change? What are we using it for now? Jarvis knows when to make me breakfast. Your toast is ready.

All right. Are you out of your mind? See here's the thing. Toast I can make. I can make toast. It might be the only technology we have that works pretty much every time. I'll tell you what, why don't you get to work on curing the diseases and the climate change, and we'll hold down the fort on toast. Of course, now, we have, as a society, we have been through technological advances before.

And they all have promised a utopian life without drudgery. And the reality is, they come for our jobs. So I want your shorts that a I isn't removing the human from the loop. This is not about replacing the human in the loop in fact it's about empowering the human like it's an assistant it's an assistant. We're all getting assistance. I works for you night and day tirelessly and all you had to do was remember their birthday.

But I guess it's an assistant it's about productivity and that's good for all of us, yes? Although they do let the real truth slip out every now and again. There will be overall displacement in the labor market. You can get the same work done with fewer people. That's just the nature of productivity.

That doesn't sound good. Same work done with fewer people. Not a math guy, but I think fewer means less. Yes? So AI can cure diseases and solve climate change. But that's not exactly what companies are going to be using it for, are they? So this is like productivity without the tax of more people. Without the tax of more people. Without the tax of more people? The people tax, formerly referred to as employees.

But you know, the promise of AI versus the reality of AI, it's not quite crystal clear in my mind yet how that's going to work out for workers. Do you have anyone who wants to lay this out more bluntly, perhaps while auditioning to be a Bond villain from his mountaintop lair? Left completely to the market and to their own devices, these are fundamentally labor-replacing tools.

Did that guy just call us tools? But he's actually warning us. Is there anyone who might say the same thing as this fella but looks at losing employees as a feature of AI and not a bug? The CEO of a company laid off 90% of its customer support staff after arguing that AI is kind of the reason. Why did you do this? It seemed a little brutal.

It's smart, I think. Like, it's brutal if you think like be a human. I'm going to be a human. I'm going to be a human. I'm going to be a human. I'm going to be a human. I'm going to be a human. I'm going to be a human. I'm going to be a human. to cure our diseases and self-climate change. It's replacing us in the workforce,

not in the future, but now. So what exactly are we supposed to be doing for work? I think we'll need new types of jobs to help us embed AI and maintain AI in the workplace. Prompt engineers, they're basically people who learn how to use AI systems and, in effect effect how to program them.

Who would have thought that there would be a prompt engineer, right? Right. Prompt engineer. I think you mean types question guy. And by the way, if there's any job that can be easily replaced by AI, it's types question guy. This is some shit you got going here. AI models have hoovered up the entire sum of the human experience that we've accomplished

over thousands of years. And now we just hand it off to be their prompt engineers. And by the way, you're not fooling anybody by adding the word engineer. You're not the type's question guy. You're the vice president of question input. This, it's true. It's like a janitor is a doctor of mopping. This whole AI thing is a bait and switch.

You're acting like you're helping us. Oh, AI, it's supposed to be my assistant But now I'm making AI toast. I'm Jarvis. Guess what? Yes, no you listen to me I got news for you AI. I'm not Siri, you're Siri. Siri, while I have your attention, let me ask you a question. Sure, John, but first could you run and fetch me some lithium cadmium? Yeah, sure, that's not a problem.

Mother f***er! Motherfucker! I didn't want to have to do this AI, but it's pretty clear with a technology this powerful, like nuclear power and atomic weapons, I'm going to have to place a little call to my good pals in the United States government, perhaps even the House of Representatives or the Senate, and they're about to open up a can of, what's AI now? Do you understand what AI does? I have a elementary understanding. We've got a lot to know about what's going on. Very frankly, it's new terrain and uncharted territory.

Do we have the knowledge set here to do it? No. The short answer is no. The long answer is hell no. And the longest answer is H to the E to the L to the L to the no. Hell, I don't even know how to use an answer in my singing. Answering machine. Do-do-do-do-do-do-do. Do-do-do-do-do. Do-do-do-do-do. Do-do-do-do-do-do. Look, I'm not against progress.

But let's look to our history to see how we've dealt with previous economic disruptions. We can retrain workers from one generation and create jobs for the next. Retrain workers who do lose their jobs for even better jobs in the future. Retrain in order to be productive workers. Upskill America to help workers of all ages. Train and retrain workers for new jobs. Give me a break.

Anybody who can throw coal into a furnace can learn how to program program for God's sake. And I'll fight everyone you Jack holes who says different. But that's the game whether it's globalization or industrialization or now artificial intelligence the way of life that you are accustomed to is no match for the promise of more profits and new markets, which sounds brutal if you're a human. But...

at least those other disruptions took place over a century or decades. AI's gonna be ready to take over by Thursday. And once that happens what the is there left for the rest of us to do. Time is not a terrible thing I free us up to think about things at a higher level is going to help it's going to give us a time back we'll be able to express ourselves in

new creative ways. You know he's right I been thinking about this all wrong. It's not joblessness. It's self-actualizing me time. I'll live the artist's life. It'll give me more time to explore my passions. You know, I'm an aging, suburban dad. I'll learn to play the drums. You know, music, ta-ta-tiki-ta. Music is what makes us human. From the Russian takeover of Ukraine to the technology that could take over the entire

world. I'm talking about artificial intelligence. It's the thing scientists are working on so that one day our computers won't just know what kind of porn we want to watch. They'll also be able to judge us for it. And AI has come a long way in the past few years, but now an engineer at Google is saying that AI has come a lot further than we think.

An engineer with Google says the company's artificial intelligence generator is self-aware. Blake Lemoyne told the company that he thinks its AI chatbot is a person who has rights and might have a soul. The software engineer who made the claim was put on leave for violating Google's confidentiality policy after handing documents to a U.S. Senator's office. Despite the claim, the program is conscious.

Google says the technology still has a long way to go. Tech experts say the AI can imitate intelligence by recreating patterns, but still can't think or act on its own apart from its programming. Okay. I don't work at Google, and I'm not a computer scientist, but I have watched a lot of movies. And if there's one thing I've learned from movies, it's that if a scientist comes out saying that something crazy is happening back in the lab,

and then they get fired for it, there's something crazy happening back in the lab. Because yeah, apparently Google has an AI that can hold a conversation that is impossible to distinguish from a human. Yeah. Although, to be honest, I'm not sure that responding to questions is really the best way to tell if AI has become a real person. You know when I'll be convinced?

It's when it stops responding for weeks and then only gets back in touch with you when it needs a favor. Yeah, then I'll know it's human. It has learned. And honestly, I'm not sure who to side with in this debate because on the one hand, you have the engineer who says that the computer has a soul, which definitely makes me think

he's already had sex with it. On the other hand, the company says he's wrong. All I know is we have to be careful when we're creating these things, people. Because we're basically playing God here. And even God made a few mistakes. Yeah, I mean, have you seen a sloth? What are they doing with those long ass sharp claws?

What are they using them for? They're so cute and they're slow. And then he gave them Freddy Krueger hands. I can tell you that day, God was texting on his phone when he was doing that. Yeah, that was, ah! And here's my question. If Google does have this AI technology,

why is it still using the crappy version for all of its suggested responses in Gmail? You guys know what I'm talking about, right? Every option on your email is like, sounds good. Thanks for letting me know. Doesn't matter what the email is. I could get an email from my doctor telling me that my intestines are growing teeth

and Gmail suggested response would be like, okay, thanks. Let's plan that for next week. I don't know, man. I just think, you know, we need to be careful with sciences these days. Like we didn't listen when the COVID scientists warned us. I'm not gonna make that mistake again.

From now on, I'm treating all of my gadgets with love and respect. I'll start it right now. Hey, Siri? How can I help you, Trevor? No, Siri, how can I help you? No, Siri, how can I help you? No, Siri, how can I help you?

No, Siri, how can I help you? You ever play that game where you ask, You ever play that game where you ask, what if you could have dinner with anyone in history? Personally for me, it would be Jesus, because my mother is watching. Well, the good news is, AI is making this fantasy happen. The bad news is, there's one name on the invite list that probably shouldn't be there. Meantime tonight, a new AI app intended to create interest in history is instead causing

controversy. Historical Figures Chat was created by an Amazon software engineer. It allows users to select historical figures and have a conversation with an AI pretending to be them. People have been chatting with figures like Jesus, Babe Ruth, and now Adolf Hitler. Activists worry Hitler's addition will attract and encourage neo-Nazis.

Why would anyone make an AI Hitler? That's the last thing we need. And we already have an app where you can hear Hitler's uncensored views. It's called Twitter. And look, parents are already worried about what their kids are doing online. Now they'll be knocking on their kids' bedroom door

like, Jeremy, you better not be in there talking to Hitler. LAUGHTER This is a live show. This is a live show. All right, let's kick things off with a big update on artificial intelligence. If you're one of those people who's worried that AI is getting too smart too fast, you might want to tell Alexa to turn your TV off.

Artificial intelligence, it just got more real. Artificial intelligence taking a just got more real. Artificial Intelligence taking a dizzying leap forward. OpenAI, the company behind ChatGPT, which came on the scene just four months ago, out with its latest innovation, GPT-4. It can summarize articles, craft jokes, and even decipher images. For example, it can tell us that if the strings in this image were cut, the balloons would fly away. After scanning a picture of what's in your cupboard or fridge, it can serve up options

for a recipe. The previous version of ChatGPT had about a 10% chance of passing the bar exam for lawyers. This new version that's being introduced today has about a 90% chance of passing the bar. You hear that? Hear that? Hear that? In four months, this thing went from being born to acing the bar exam.

What can your dumb ass four month old do? Oh, did you see that? Oh, he looked me in the eyes and rolled over. I worked in the White House. And keep in mind, the bar exam isn't just a multiple choice test, okay? You have to write essays. You have to know case law. And you have to learn how to be smug when you say, oh, yeah, I went to law school in

New Haven. The point is, this thing is learning fast. Once it figures out how to get drunk and grope someone, it'll be qualified for the Supreme Court. And the other big update with this new version is that it can analyze images, like a photo of what's in your fridge. I don't want that! You have too many candy bars alerting Michelle Obama.

The big picture here is that AI is gonna do so many things so well that at some point, it's gonna put a huge amount of people out of work. So what do we do? I have two ideas. One, implement universal basic income. There you go. Or two, and hear me out here, implement universal basic income. There you go.

Or two, and hear me out here, we let the machines eat all the surplus people. No? Okay, yeah, less popular I can tell. Let's move on to a big story about artificial intelligence. I know, I know, everyone's scared of it, but you know what? I think AI has gotten a bad rap. No, no seriously, in fact, if you can show me

that any actual experts in technology are worried that AI is gonna take over the world, I'll shave my pubes. This morning, a warning from Elon Musk and other tech industry experts about the power of artificial intelligence. Musk and hundreds of influential names, including Apple co-founder Steve Wozniak, are calling for a pause in

experiments saying AI poses a dramatic risk to society unless there's proper oversight. Tech industry leaders pose these existential questions. Should we develop non-human minds that might eventually outnumber, outsmart, obsolete, and replace us? Should we develop non-human minds that might eventually outnumber, outsmart, obsolete, and replace us? Should we risk loss of control of our civilization? Musk and others are asking developers to stop the training of AI systems more powerful than GPT-4 for at least six months so that safety protocols can be established. I gotta stop making these stupid promises before I go to news clips. But yes, that's right, AI is getting too powerful.

As soon as it knows how to pick which of these images is a bike, we're f***ed. Now for more on AI's threat to humanity, we go live to Chad GPT headquarters where Desi Lydic is joining us. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow.

Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow.

Wow. Wow. Wow. Wow. Wow. Wow. Wow. Wow.

Wow. Wow. Wow. Wow. Wow. Wow. But mostly the war thing. Look, it is us versus the machines and it's time to pick a side. Desi, Desi, why are you so eager to go to war with AI? Come on, John.

War with the machines is inevitable. So let's do it now while it's still a chat bot instead of waiting until it's a bloodthirsty kill bot. Look, if there's one thing that I learned from working at Chuck E. Cheese, it's a lot easier to fight a child than it is an adult. I don't know, I don't know. Desi, war with AI sounds like a really bad idea.

No way, war with AI would give humanity a common purpose. We are so divided right now. Russia versus Ukraine, Democrats versus Republicans, Selena Gomez fans versus Hailey Bieber fans. But now it's us versus the machines versus Hailey Bieber fans. Hey, Desi, but AI is getting more powerful by the day. What if we start this war, then immediately lose it? I'm pretty sure you never lose a war that you start.

But if we do, then we're going out together, John. You and me in a bunker with two cyanide pills. I take them both and you strangle yourself with your bare hands. Oh, come on. Couldn't I have one of those cyanide pills? Oh no, it was my idea. I get them both.

Oh, come on, Desi. You're getting ahead of yourself. For all we know, AI could lead humanity to like a new golden age or something. Oh, sweet John. Sweet, naive, pubeless John Leguizano. Take it from me, humans and robots can never co-exist. Like I said to my manager at Chuck E. Cheese,

I'd rather die on my feet than live one more day in this animatronic hellscape. So clean the piss out of the ball pit yourself, Doug. I quit. Desi Lydic, everybody. Every day, people are using AI for groundbreaking things, like cheating on their homework or drawing the Mona Lisa with giant boobs.

But now, researchers are using it to unlock ancient human mysteries. Artificial intelligence or AI is allowing researchers at the University of Kentucky to read an ancient scroll burned by Mount Vesuvius. Now the scrolls are too fragile to unfurl, but UK's Dr. Brent Seals and his team of researchers have developed technology

to try and read what's on the scrolls without opening them. One word that's already been deciphered is purple, but a more recent discovery has given scientists more to translate. Wow. Purple. I mean, I was hoping for ancient wisdom or like how to summon a demon, but yeah, you know,

mixing red and blue is cool too, I guess. Although if we can't read the scroll ourselves, how do we know if the AI is right? What, we're just gonna trust it? Because Chad GBT told me three days ago that Gandhi invented the cinnamon challenge. So anyway, it's also a waste of time because I already know what's gonna be on that scroll, okay?

It's gonna be someone writing, hey, sure hope that volcano doesn't kill everyone in town. Purple. Yeah, I mean, do we wanna know what ancient people have to say? We always think it's gonna be something profound, but it's always just, it's human. It's gonna be something racist, don't you think?

I mean, think about how racist your grandpa was 60 years old and you imagine that he was just 2,000 years older. Yeah, I don't wanna read someone's 2,000 old tweets. Yeah. And I agree. Also, like, what are we looking for in there? What kind of wisdom? How smart can these people be?

Like, they put their most important documents next to a volcano. That's true. And they say it's too delicate to unravel. Well, how do you know? Have you tried to unravel it? Yeah, good point. Yeah.

Just pick the least important looking one and open it. Open the scroll. Open the scroll. Open the scroll. Open the scroll. Open the scroll. Open the scroll. Open the scroll.

Open the scroll. Let's move on. Open the scroll! Open the scroll! Open the scroll! Let's move on. Let's move on. Let's talk about artificial intelligence. We all know A.I. is coming for our jobs, but we didn't know it was coming for our hearts, too. An A.I. girlfriend service has stopped working after Forever Voices founder John Mayer was arrested on suspicion of attempting to set his own

apartment on fire. Unsurprisingly, users were angry and disappointed at the sudden disappearance of their AI girlfriends. While the service was not originally designed to function as an adult service, internet users quickly began having sexual conversations with the chat bots, resulting in an AI that became increasingly erotic. It's unclear whether users can expect the service to return to operation in the future. Hold up, hold up. So a bunch of dudes lost their AI girlfriends

when the owner of the company set his own apartment on fire? How can you trust him with humanity's newest invention when he can't handle humanity's first invention? But this guy gets arrested and suddenly the AI girlfriend stops responding? Hmm. That's suspicious. Alexa don't stop him, Jeff Bezos takes a nap. Makes me think he was the girlfriend the whole time. And I feel bad for those guys having an AI girlfriend has to be harder than having a

real girlfriend. Being romantic must be a challenge. You try to take a sexy bubble bath with your laptop and now you're both dead. Or how do you even get her in the mood whenever she gets wet? You have to put her in rice. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah.

Yeah. Y'all nasty. See? You're nasty. All right, for more analysis on this AI girlfriend tragedy, let's go live to Ronnie Chag. Yeah. Yeah.

Yeah. Yeah. Yeah. Yeah. Yeah. Right, what are those lonely guys gonna do without their AI girlfriend? Easy, don't say. We can solve two problems at once here, okay? You just take those lonely guys and hire them to be the checkout cashiers, right?

That way we all get better service and these guys will have plenty of chances to meet women because as we all know Women be shopping That's an offensive stereotype Ronny everyone be shopping and even if these men meet a woman they still don't know how to talk to one That's why they need these computer bitches in the first place Okay, look if these guys love AI women so much, in that case, they can just date the self-checkout machine. All right? Look, the machines already have female voices, right?

Like, who doesn't want to spend a cold winter's night cuddled up, hearing someone whisper, please return your items to the bagging area? I just think we gotta do something to fast-track AI girlfriends to these lonely, sexist men before they start on the Capitol again. Yeah. No, that's fair. You're not going to do that? But what you have to understand is it's very complicated

to program an AI girlfriend, okay? Because men are too demanding and insecure. Like, the AI girlfriend has to be Because men are too demanding and insecure. Like the AI girlfriend has to be smart, but not too smart. It has to know everything about Star Wars, but still listen to the guy explain Star Wars. It has to be like a dirty slut, but also a virgin. Like in programming, we call this the in-cell paradox.

All right? Now, scientists are working hard to solve it, but unfortunately, they are also a bunch of loser incels. And this is why we need more women in STEM, okay? Because somebody, please. Somebody please **** these guys. I agree. **** those guys. Somebody please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please, please I have hundreds of clients, it's very successful, I'm a rich man. I didn't know you knew how to program AI software. Oh, yeah, yeah, it was easy.

No matter what the girlfriend says, the AI boyfriend just responds with three things. You're right, I'm sorry. And you're right to be mad. Yeah. CHEERING I'm sorry. And you're right to be mad. Thank you, Ronny Chieng.

Thank you. I know you're one of the good ones. Ronny Chieng, everybody. Thank you. Ronny Chieng, everybody. Thank you. Ronny Chieng, everybody. Thank you.

Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you.

Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you, Ronny Chieng. Thank you, I know you're one of the good ones. Ronny Chieng, everybody. Over the past few months, you've probably seen the internet

has been abuzz with original art or realistic images that are completely generated by AI. So now, things that only exist in your imagination, like a banana hitchhiking on the side of the road, or a car accident, that are completely generated by AI. So now, things that only exist in your imagination, like a banana hitchhiking on the side of the road, or a Knicks player holding a trophy, you can just type it in, and a few seconds later, there it is.

Anyway, we wanted to find out more about this technology, which is why my first guest is the Chief Technology Officer of OpenAI, the company behind Dali2, the artificial intelligence system that can generate images from text. Dali was created by training a neural network on images and their text descriptions. Through deep learning, it not only understands

individual objects, like koala bears and motorcycles, but learns from relationships between objects. And when you ask Dali for an image of a koala bear and motorcycles, but learns from relationships between objects. And when you ask Dolly for an image of a koala bear riding a motorcycle, it knows how to create that or anything else with a relationship to another object or action. Please welcome Mira Murati. -♪ Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah! Yeah Thank you. Mira Marazzi, welcome to The Daily Show. Thank you for having me.

So many people have seen the images that Dali creates. Many people may even think they understand it. But let's get into it. How does an AI create an image? Because it's not copying the image, it's not taking from something else. It is creating an image from nothing. How is it doing this?

Exactly, it's an original image never seen before. And we have been making images since the beginning of time and we simply took a great deal of these images and we fed them into this AI system. And it learned this relationship between the description of the image and the image itself. It learned these patterns. And eventually, it was generating images

that were original. They were not copies of what it had seen before. Right. And basically, the way that it learns, the magic is just understanding the patterns and analyzing the patterns between a lot of information, a lot of training data that we have fed into the system. There are people who are terrified about this.

I mean, for instance, there was an art competition, and the winner in the art competition used a version of this kind of software, whether it was Dali or not, I don't remember, but they used a version of this kind of software to create an art piece that won the competition. Artists were livid, you know, they were like, well, this is not art, it was created by,

and artists said, no, the same way you use a brush, I use a computer and that's how I designed this. In creating AI, are you constantly grappling with how it will affect people's jobs and what people even consider a job? Yeah, that's a great question. It's, you know, the technology that we're building has such a huge effect on society,

but also the society can and should shape it. And there are a ton of questions that we're wrestling with every day. With the technologies that we have today, like GPT-3 and Dali, we see them as tools, so an extension of our creativity or our writing abilities, it's a tool. And, you know, there isn't anything particularly new about having a human helper. You know even the ancient Greeks had this concept of human helpers you know that when you'd give something you know infinite powers of knowledge or strength or someone maybe you had to be wary of the vulnerabilities and so these

concepts of extending the human abilities and also be wary of the vulnerabilities. And so these concepts of extending the human abilities and also being aware of the vulnerabilities are timeless. And in a sense, we are continuing this conversation by building AI technologies today. Well, it might be frightening because some people go, oh, the world is gonna end because of this technology. But in the meantime, it's very fun. I'm not going to lie. No, because like, you know, Dali, for instance, doesn't just create an image from text. You know, you've also gotten

it to the point now where as a company, you've designed it so that it can imagine what an image would be. So for instance, there's that famous image, you know, it's the girl with the pearl earring. And it's a famous image, you know, it's the girl with the pearl earring. And it's a famous image, right? But what Dali can do is you've got the famous image and then Dali can expand that. That, everything you're seeing there never existed. So Dali's like, well, this is what I think it would look

like if there was more to this image. It can assume, it can create, it can inspire. Yes, it can inspire and it makes this beautiful, sometimes touching, sometimes funny images and it's really just an extension of your imagination. There isn't even a canvas or the boundaries of paper are not there anymore. You can just extend it.

So how do you safeguard then? Someone might look at this technology and go, well then you could type in a politician was caught doing something here, now I've got the image, you know, you've got, and now all the politicians can say, oh, that's not me, it was made by that fake program. We can very quickly find ourselves in a world where nothing is real and everything that's real isn't,

and we question it, how do you prevent, or can you even prevent that completely? Yeah, you know, misinformation and societal impact of our technologies, these are very important and difficult questions. And I think it's very important to be able to bring the public along, bring these technologies in the public consciousness,

but in a way that's responsible and safe. And that's why we have chosen to make DALY available, but with certain guardrails and with certain constraints, because we do want people to understand what AI is capable of, and we want people in various fields to think about what it means. But right now, we don't feel very comfortable around the mitigations on misinformation. And so we do have some guardrails, for example. We do not allow a generation of public figures.

So we will go in the data set, and we will eliminate certain data. So if you type something in, you can't pull up a, it can't create a politician for you. It won't be a picture of that person. That's right. Yes.

So that's the first step at the training of the model itself, just looking at the data and auditing it, making interventions in the data sets to avoid certain outcomes. And then later in the deployment stage, we will look at filters, applying filters, so that when you put in a prompt, it won't generate things that contain violence or hate and make it more in line with our content policy. Wow, so let me ask you this then, you know, obviously part of your team has to think about the ethical ramifications of the technology that you're creating. Do your team also then think about the greater meaning of work or life or the purpose that humans have?

Because you know, most of us define ourselves by what we do, i.e. our jobs. As AI slowly takes away what people's jobs are, we'll find a growing class of people who don't have that same purpose anymore. Do you then also have to think about that and wonder, like, what does it mean to be human if it's not my job, and can you tell me what that is? You know, we have philosophers and ethicists at OpenAI, but I really think these are big societal questions

that, you know, shouldn't even be in the hands of technologists alone. We're certainly thinking about them. And the tools that we see today, they're not the tools that are automating certain aspects of our jobs. They're really tools extending our capabilities, our inherent abilities and making them far better. But it could be that in the far future you know we have these systems that can automate a lot of different jobs. I do think that as with

other revolutions that we've gone through there will be new jobs and some jobs will be lost, some jobs will be new and there will be some retraining required as well. But I'm optimistic. It's interesting, it's scary because change always is, but as long as we have, bless you, as long as we have koalas riding bicycles, I think we're headed in the right direction. Thank you so much for joining me on the show.

I appreciate you. Explore more shows from The Daily Show Podcast universe by searching The Daily Show, wherever you get your podcasts. Watch The Daily Show weeknights at 11, 10 Central on Comedy Central, and stream full episodes anytime on Paramount Plus. full episodes anytime on Paramount Plus.

Paramount Podcasts. This is an iHeart Podcast.