(keyboard keys clacking) (dramatic orchestral music) - Suddenly, it seems everyone is talking about artificial intelligence. - Well, well, let's take a look at a video that never happened, because OpenAI, the company behind ChatGPT, has unveiled a new artificial intelligence model that takes written prompts and video. - Self-driving robotaxis are already on the road in two US cities and could soon be coming to a place near you. - My books have been taken, and I didn't even know about it. I wasn't asked about it. I didn't approve it. - The violation of me, of my individuality, my likeness, my voice are saying these words that I did not agree to say. - Who's that guy? Why did they all stop and run into each other? - The AI that you are most likely to encounter in your life is built by a handful of for-profit companies.

using data you never realized they were collecting. (dramatic orchestral music) - [Eddy] So, a few months ago, I'm wasting precious hours of my life away on TikTok when suddenly I'm shown an ad for Google's AI program, Google Gemini. - Okay, it can't be that hard. ♪ How about generate an image of a cat playing guitar ♪ ♪ Is that how it works? ♪ ♪ Am I doing AI? ♪ ♪ Yeah, it just does whatever you type ♪ - [Eddy] This was the first commercial from a big company I'd ever seen for AI, and I was surprised by the lifestyle it pitched. ♪ What are last minute gift ideas ♪ ♪ You can make with arts and crafts ♪ ♪ Plan a workout routine to get bigger calves ♪ ♪ If a girl calls me a snack ♪ ♪ How do I reply ♪ ♪ Yeah, that's how it works ♪

♪ You're doing AI ♪ - [Eddy] Rather than taking care of things you don't wanna do, like organizing your schedule or your job, the pitch seems to be that it would help you think about your loved ones or communicate for you. But it begs the question: why do I need to use this? Couldn't most of this just be done by Googling something and finding sourced information? Is AI in its current state going to sincerely improve our lives, or do large corporations want us to adopt it just because it's the hot new thing? ♪ What's a good excuse to cancel dinner with my friends ♪ - We're literally sitting right here. ♪ There's no wrong way to prompt ♪ ♪ Yeah, you're doing AI ♪ ♪ There's no wrong way to prompt ♪ ♪ It does whatever you type ♪ ♪ Just prompt your prompt in the prompt bar ♪ - And since that ad,

The floodgates have burst open with AI advertising. I now constantly see AI commercials from Meta, Google, and other companies on TV. My neighborhood has been invaded by self-driving Ubers. So you know what? I wanna know what life looks like when someone tries to apply this technology in its current state in as many ways as possible. But I take a while to make videos, so let's get back to earlier this summer. Over the last year, we've all been constantly bombarded by new AI programs. We've got ChatGPT from OpenAI, Meta AI from Meta, Google Gemini, and many more. During this time, it seems like anything that a computer does is now being branded as artificial intelligence. But one of the main introductions and advertisements for AI is its ability to generate images. ♪ Generate an image of a cat playing guitar ♪ - Imagine a tree house- - [Child] In a big oak tree with a porch and a ladder.

- Imagine Little Italy in 1954. - Oh, there it is. - [Child] Wow, it's amazing! (father laughing) - Anything that your heart desires. Do you wanna see a super cool guy? Well, yeah, I guess this is what Google Gemini gives you. Now, I'm headed to Chicago for the weekend, so I thought I'd ask Gemini to generate images of the city before I go. Now, it gave me four images, three of which are weird, messed up photos of the Chicago skyline. I don't know what this is supposed to be. The fourth image was just this photo of a deep dish pizza. I said, Generate images of Chicago, and it gave me this as the fourth one. Now, we all know deep dish is famously made with its crust, tomato on top of the crust, cheese on top of the tomato, tomato on top of the cheese,

and then cheese on top of the tomato with pepperoni on top of the cheese. Sounds like a regular deep dish pizza to me. Once we were there for the weekend, we started off by grabbing a picnic blanket, beers, and watching some 4th of July fireworks, one of the most classic ways to enjoy the summer as an American, but could it be improved on? The crowd was focused on the dazzling light show in front of us. (patriotic music) (fireworks crackling) Drew kept asking me to look up and get off my phone, but I refused to look, because I had AI-generated images of fireworks myself, so I didn't feel like I missed out on anything, actually. It was like I had my own private show on my phone. It was actually, it was just as good. The next day, we headed downtown and went on Chicago's architecture tour. This is usually my number one recommendation

for anybody visiting the city for the first time. Chicago's architecture is absolutely beautiful. Each building comes with its own artist and story of how it was constructed, when it fit into the city, what was happening at the time. Thousands of human hands and lives touched each individual building. Every detail on every building was individually constructed and crafted by human choices. These things make the skyline a beautiful mosaic of art, architecture, engineering, construction, and a history of over 100 years, all of these aspects and human lives beautifully swirling down a funnel that feeds an AI image model that can spit out its own version of the skyline with random new buildings that have no stories, that have no individual touch to them. But they popped up on my phone, and it's almost like I made them. Finally, we ended the weekend at the Lincoln Park Zoo. (people chattering) (birds chirping) Yeah, cool rhino, idiots. The one I generated is freaking on fire.

So obviously, AI image generation can't compete with real life, but what are the advertised uses for this thing right now? There were two in the Meta ads shown earlier that I wanted to call attention to. First, a father and daughter planning a tree house. They don't ask AI to analyze a photo of the tree and pick out a structural blueprint for a tree house that would fit perfectly. They just say, Imagine a tree house in a big oak tree with a porch and a ladder, which sure, I guess it could help you like pick out a vibe, but you need a ton of research and sourced information to plan a tree house that's safe for your child. But the ad just kind of shows that they built it. They looked at this photo, and then they just built a tree house. You can also just Google information about a tree house in a big oak tree, and not only will you get

people's experience and planning, and information that'll help you. You also get visual representations of what it could look like. But the one I have a much bigger problem with is the other example in the Meta commercial. Here, a younger man, assuming a grandson, asks Meta to, Imagine Little Italy in 1954. AI image generation in its current state gives you a fictional image based on visual and contextual information that it's been fed. So when you're looking for a photo of a real time in history, you should not be using AI. Here's a photo of Little Italy in 1974 taken by Eddie Hausner for The New York Times. There is so much information to learn just from one photo of the past. Just on a surface level, you can see how someone worked, what the fashion in the area was like at the time, items that stores were selling, what graphic design and signs looked like at the time,

building architecture, hell, even what awnings looked like for all my awning-heads out there. Real information about a real moment in time with real people captured by a talented photographer. This is not Little Italy in 1954. This is just the vibe a computer thought Little Italy was like. - Wow, there it is. - Even when used as a building block for an idea, AI image generation is an amalgamation of stolen art and information. Using it as concept art not only steals from actual artists, but it robs you the experience of finding an artist you really like. It robs your inspiration of personal human touch. So, okay, I don't really like AI images, but I wanna try something that exists in the physical world, a nearly 5,000-pound vehicle that thinks for itself, the Waymo car. - Well, as if the roads and the highways are not clogged enough, you might see even more vehicles on the roads.

That's because Waymo is testing its self-driving cars on San Francisco freeways. - Startup Waymo moving forward with expanding its robotaxi service here in L.A., despite ongoing concerns over the safety of these autonomous vehicles. - These were originally Google's self-driving autonomous vehicle in 2009 until it was spun off and established as Waymo in 2016. And because of my proximity to them, I can call one right now, get in the car with my physical body, and take a deep exhale, really breathe outward into the car. Uh... After my COVID quarantine and a negative test, I was ready to stop spewing my COVID germs, and go back to spewing my regular germs in a Waymo car. But first, a quick word from the sponsor of this video. I fell down a well. Don't ask. This video was sponsored by CookUnity. CookUnity is the first chef-to-consumer platform.

delivering freshly prepared preselected meals right to your door weekly, which sounds tasty. I could really use some food down here. CookUnity chefs offer up a wide range of meals with over seven different dietary preference filters, including vegan, paleo, and gluten-free options. The meals are delivered fully cooked, so all you have to do is heat them up, and their subscription service is super flexible, so you can pause, cancel, or skip weeks at any time. They sent me a few meals, including Southern Fried Chicken with Spicy Honey and Buttermilk Mashed Potatoes by Chris Ratel from New York. I also got an eggplant parm dish, some chicken katsudon, and a couple more. All of the dishes were not only easy to cook; I just popped these bad boys in the oven for around 15 minutes, or whatever the directions said, but they also did taste great. So go to cookunity.com/burback50,

or click the link in the description, and use my code BURBACK50 to get 50% off your first order of CookUnity meals to try them out for yourself. Okay, back to the video, but just, you know, before we cut away, if anyone could call for help, or get me out of the well. Now, the current testing area for these cars stretches from Downtown L.A. to Santa Monica. So I decided to go for a random round trip just to see what the ride was like. (car engine humming) The car was a white Jaguar SUV, and it came with its own self-driving hat. But as I reached for the handle... Wait, how do you open this? I could not figure out how to open it. Genuinely, how do you open the car door? There was a fruit cart on the corner, and the person selling fruit absolutely laughed at me. It was like a loud laugh,

like they did not care that I heard them laughing at me. But it turns out you need to go on the app and unlock the car door first. So once I did that, the handle popped out. - [Car] Hey, Eddy. (car door thudding) - [Eddy] Hello. Once in the car and buckled up, I was ready to take my first ever self-driving taxi. (light music) - [Announcer] Hello from Waymo. As we get going- - What? - [Announcer] Please give us one minute to cover a few riding tips. This experience may feel futuristic, but the need to buckle up is the same as always, so keep your seatbelt fastened, please. - [Eddy] Other than being very uncomfortable with the fact that there was no driver in my car... So how long have you been driving for Waymo? Things were going pretty smoothly. With the limited range of where Waymo cars can go

in the city right now, I decided to have it take me to the Beverly Center. I was doing a round trip, so I wouldn't leave the car. But if you've lived in L.A. for a long time, you'll know that the Beverly Center is a meeting place of culture, a spot where small businesses can thrive. Anytime anyone I know visits L.A. and they ask for spots, I give 'em my top five: Santa Monica Pier, the Beverly Center, Hollywood Boulevard, The Grove, and The Americana. It's a rite of passage for L.A. transplants to go to small, wealthy, boring areas of the city, and then go on social media to say that all of L.A. sucks. Anyways, back to the ride. So... When did you start driving for Waymo? We made it to the Beverly Center drop-off without any issue, and since it was a round trip,

I had the option to just resume the ride again. As I went to resume the ride, I scared myself because there was a piece of lint on my pinky ring, and I thought it was a bug. Ooh! After being literally so brave and getting the piece of lint off my pinky ring, I resumed the ride. However, we were in a tight parking lot, and the self-driving car decided to stop driving itself. (tense music) Okay, so- (console chiming) So there's just a simple- (console chiming) We sat there stuck between an AI rock and a completely open place. (tense music) That's it, we just gotta turn out. There was also a parking booth right in front of us, and this guy seemed so frustrated because he knew it was a self-driving car, and it wasn't moving at all. But after a moment, it seemed like support stepped in, and we started moving just fine.

Just go! Oh, he's looking at me. But other than our slight delay, the rest of the ride was pretty smooth and comfortable. The driving these things have to do is relatively easy. There's no highways or anything, but this does make me nervous for the ride-sharing industry, especially because at the beginning of this ride, I did price match it with Uber, and at least at that moment, it was $5 cheaper. Do I think this new startup will be cheaper forever? Yeah, I'm willing to bet my MoviePass on it. But Waymos haven't exactly been a smooth ride all over the country. They're currently testing in L.A., Phoenix, and where they originally started, San Francisco. Now, to the driverless cars that some San Francisco residents say are making them lose sleep with honking horns all night long as they try to park themselves. Larry, back to the story, I'm glad none of us live there.

We'd have a problem. (car horns honking) - [Newscaster] This morning, San Francisco residents are fed up after being kept up at night by driverless Waymo cars honking at each other through the night as they park. (car horns honking) - [Sophia] It's fucking 4:00 a.m. - Past two weeks, I've been woken up more times overnight than I have combined over 20 years. - [Eddy] It's also particularly shocking to me that they can just be on the road testing when it feels like we don't have a lot of rules for that yet. There were multiple incidents of self-driving cars being pulled over by the police, and they just like don't really know what to do. - [Newscaster] An officer gets out of his car to speak to the driver, only to discover- - Ain't nobody in it. - [Newscaster] And then the car speeds away.

from the officers, passing through an intersection before pulling over again. - [Police Officer] No, can you show me out at 7th Ave and Osborn? I got an unmanned vehicle that was kind of taking off from me. It's a Waymo car. - [Newscaster] It's not every day you see a Phoenix police officer pulling over a driverless car. - [Police Officer] Hi. - [Eddy] But with all that being said, my ride was somewhat smooth, and it was coming to an end. But yeah, after a few minutes, I got to the end of the ride, and it was time for me to get out. (light music) (car door thudding) But as I entered the sight lines of the Waymo rear-view mirror, I realized that it wasn't just my first autonomous taxi driving away. No, on that fateful summer, I had made a friend. ♪ Best friends make better lovers ♪

♪ Best friends make better lovers ♪ ♪ Best friends make better ♪ - [Eddy] How long have you been driving for Waymo? ♪ Baby, baby, yeah ♪ ♪ Best friends make better ♪ - [Eddy] Okay, so- (console chiming) ♪ Me and you ♪ ♪ Best friends make better love ♪ - [Eddy] Goodbye, Waymo. But Waymos aren't the only self-driving vehicle in my neighborhood. (upbeat music) - [Newscaster] Now at 5:30, Hollywood's food delivery robots are under attack. Videos show vandals kicking them and even tipping them over. - [Eddy] Food delivery robots, you may have seen them getting beaten up on the news, driving through an active crime scene. (people chattering) - [Bystander] Hey, how come the robot can go over there? - [Eddy] Or trying to cross the road on a red light.

This is not my video, but I have seen this exact thing happen before. I watched one go out into the road on a red light; a car had to screech to a halt and honked at it, which, like, I mean, the robot's not gonna be like, Oh my God, I'm so sorry. If you're unaware, around the time of the pandemic, these little guys started driving around L.A. They show up to a restaurant, lock the food in their little box body, and then drive it over to your address. But I mainly know these robots from them getting in my way while I'm on a run. (people chattering) This one was recorded just a few days ago while I've been making this video. It almost ran into me, and after I finished my run, I circled back about 10 minutes later to see the progress it had made, and it had only moved one block.

These things just sit there, not move. Oh, it's actually going. Oh, I stand corrected. It is on its way. Oh, nope, it's on its way. It's going right to the address. Oh, nope. Okay, nope. Okay, now it's going, and I think it's gonna make its way to the address just, oop. And now it's going in the grass. It got stuck in the grass. But there's something glaringly obvious when you look at both of these self-driving robots, the main reason for them existing that we can't ignore. These are here to replace human work. So assuming nothing goes terribly wrong with the Waymo cars and delivery robots, these things may be the perfect example of the real crisis we'll be facing in the future. New technologies and labor have always done two things at once. One, it makes the task of labor easier to complete, sure,

but two, that easy labor comes with less of a need for the number of laborers. Let's take the taxi business, for example. It was all but completely destroyed by ride-sharing apps like Uber and Lyft due to convenience. Now that we have millions of ride-sharing drivers in the US alone, what happens to them if Waymo is a success? Now, I can already hear the popular argument for technological advancements in work. Well, okay, sure, people will lose their jobs, but that technology will create new jobs. Machines need engineers, people to produce and build them, people to repair them. But with the evolution of technology, and specifically the aim of AI, its convenience by definition would require less and less human work. Let's say Waymo puts all ride-sharing drivers across the United States out of work. Are they all gonna become AI programmers? AI progressing will just take those coding jobs too, no?

So my question is, if AI gets better, what is the plan here? Even if these people found a different job, how could you guarantee the different job wouldn't also be eaten by AI? Isn't that the whole goal of making AI more intelligent? And in fact, even on the consumer end, I've been taking a couple of Waymos over the last few weeks. How long you, how long you've been driving for Waymo? Right, sorry. And there's even been times where it is more expensive than an Uber without any human labor involved. There's no driver, and yet, it's somehow more money? So what's the point? The point would be not to make it cheaper, or more convenient for us, but not having to split that profit with a driver that works for you. It's the same, if not more out of your pocket as the customer, but the whole pie goes to them, and that's considered smart business.

until no one has any damn money to spend. I mean, this type of thing would require a completely new system for working and living. Like even if most corporations had AI take 99% of the jobs in their companies, how the hell are they gonna make a profit if 99% of people are out of work? Right now, the Beefy 5-Layer Burrito at Taco Bell is $4.59. It was 89 cents in 2009. 89 cents in 2009 calculated for inflation today is $1.30. $1.30 is not over $4 last time I checked. And if a computer digitizes a 5-Layer Burrito into existence without any human work, are you gonna make it cost way less, Taco Bell? Even if it's cheaper, what money will I have if AI takes my job? My job isn't hard. What money will any of us have? This is the most shortsighted bullshit I've ever seen.

Even if you're mega wealthy, you can't keep making gains on your investments if nobody has any damn money to spend. Why would you want computers to take all our jobs? And I don't even know if AI is that good, or will be that good, but what are we thinking? All right, so maybe I got a little too heated about the Beefy 5-Layer Burrito. So I decided to use AI in a way that usually calms me down, watching movies. And the big scare for movies right now are AI video models like Sora AI and Luma Dream Machine. - It's an AI technology that generates video from a text prompt. So you describe what you want to see briefly in a sentence, or a paragraph, the system will generate what you've described. - [Eddy] I'll be honest, I've been pretty closed off to the idea of AI being used in movies in any way.

It seems like while some of it can be used as a tool, using it to replace creative work will take out the richness, detail, and hard work that's poured into every frame. But maybe I'm just afraid of change. So I decided to get cozy and sit down to watch what AI video has to offer. First, I found something called Star Wars- 1950s Super Panavision 70 AI Film. - [Narrator] In a distant galaxy where the eternal struggle between light and darkness plays out among the stars, follow Luke Skywalker, a young farm boy destined for greatness, alongside Princess Leia. Opposing them is Darth Vader, the formidable Dark Lord of The Sith. - [Eddy] Okay, so it's just shitty AI GIFs that look much worse than stills of the actual movie narrated by an old-timey AI voice. Not a great start, but what the hell is The Simpsons by Adam Sandler Trailer Comedy Movie?

- [Homer] Hi there, I'm Homer Simpson. Welcome to our crazy little world. Let me introduce you to my family, friends, and some of the other colorful folks in Springfield. First up, there's my lovely wife, Marge. She's the glue that holds this family together with her patience and love, and check out her hair. Isn't it something? Next, we've got my boy, Bart. He's a little troublemaker, always getting into mischief, but deep down, he's got a heart of gold. - This fucking sucks. - [Homer] And there's Lisa, our little genius. She's the smartest one in the family by a mile, always reading, always thinking, and always trying to make the world a better place. Now, there's Ned Flanders. - This isn't a trailer for a movie. It's just disturbing shots of who they'd cast in the live-action

while a shitty Adam Sandler voice just describes everyone. - [Homer] He's a bit too perfect sometimes, but you gotta love him. We've got my buddies from work, Lenny and Carl. We share a lot of laughs at the power plant. - And it just keeps going. What movie trailer have you seen that's like, Hi, I'm Iron Man. I'm a pretty smart guy. I'm also friends with Captain America, who's really strong, but not as strong as The Hulk, who's a big green guy. Hawkeye's here, we don't really know why, but he's got a great app. The team is held together by Black Widow, who for some reason is married to Colin Jost, and I still don't get it. I mean, sure, Colin seems like a nice and funny guy, but really, those two? For real?

Hold on, this was uploaded 13 days ago and has over 2.7 million views. I think these are really scary for the CGI industry and the art that goes into that, but this is nowhere close to even making scenes where two people can talk to and react to each other. As I continued browsing, everything I watched, no matter the genre, had one thing in common. They were all deeply, deeply sauceless, just boring, uninspired, mashed-up collections of other things I'd seen before. There was no vision, no direction, no expression involved, like this one, Genesis, this movie entirely made by AI, 4K. - [Narrator] In a fleeting heartbeat of time, humanity's own creation sparked a nuclear inferno, turning our vibrant Earth into a solemn testament of our own folly. The world as we knew it was engulfed in nuclear fire,

marking the end of an era. The sole survivor clung to a vision, a blueprint for forging a new future from the ashes of the old. I feel so conflicted about these, because yeah, it is crazy to see an AI create these visuals when they didn't do that, you know, like months ago. But as somebody who grew up on and absolutely loves entertainment, whether it be, you know, games, movies, television, what the fuck is this? If you wanna look me in the eyes and tell me that something like this is comparable to Blade Runner 2049, I'll throw my head through a wall. Are you serious? There's not even a scene here. There's not even a character you can follow. Everybody looks different in every shot, and everything we're seeing is just a weird mashup of every sci-fi thing it's been fed. There's no planning. There's no cool stories from the set,

or the creation of any of the ideas. It's just stuff. You know, a few nights ago, I was watching an Italian movie called Cinema Paradiso. The film follows a young boy who falls in love with the movie theater in his small town in Sicily. The director of the film actually shot in his hometown as a love letter to people in his life and what the art form of filmmaking has given to him, and it's not just a coming-of-age film based on his childhood. You can tell that as an artist he's returned to his roots to realize just how important people who guide you through life can be to you, even many years later. My explanation is not doing the film justice, all right? I'm not a good writer, and the movie is incredible, but that's how real artistic expression works. Any good sci-fi isn't just made to be a sci-fi movie. A collection of people, all in their own ways,

are using their life experience and the experiences of those around them to express themselves, whether that's in editing, costume design, directing, acting, writing, set design, sound. When a movie really lands with you emotionally, it's not some coincidence where an algorithm hit just the right numbers to make you cry. While searching for AI films, I did find a super short one called Poof. It's not really anything, just a series of Muppets exploding at work, but it seemed a little bit more creative than these weird computer mashups, which made sense when I looked into it, and there was a creator who at least used AI to compile a lot of aspects together to make the thing. And whether I like it or not, I'm sure using it as a tool will become more common. But I still think the more you let computers handle the art part completely on their own, the less people will connect with it.

Cinema Paradiso, which I highly recommend you watch, made me cry, and I'm very grateful that everyone involved with that movie expressed themselves the way they did when they did and worked together to create it. I think that kind of collective human art is really special. So yeah, I think the way as a human being that movie spoke to me is gonna scratch the exact same itch as Adam Sandler's AI The Simpsons. Also, I should add that if there was a sentient AI, of course, I'd be interested in the art that it would make. If it worked like a human brain, but was completely different, I would be totally interested. But if you tell me that this shit is gonna replace human expression and human art, get fucking real. I'm gonna see Kristen Wiig's Marge in my nightmares tonight, and maybe in my dreams if I'm lucky. Sorry, sorry, I got, let me go apologize to my girlfriend.

The night I recorded this, I was walking to get some drinks with friends, and I had this feeling that I was being followed, and then I saw it. On my tail was a delivery robot following me for multiple blocks. It was sure to never get too close, until finally it made a pass at me. (cart wheels rattling) (Eddy laughing) This is how I know I'm doing important work. The robots are getting mad at me. Now, I'm sure that if you weren't exposed to AI with image generation, it was most likely through hearing about the text-based AI ChatGPT. - [Newscaster] Ammaar Reshi asked a computer program to write a book. - I think it was, Write a children's book about a young girl who creates her own AI. - [Newscaster] And in a weekend, Alice and Sparkle was finished. - [Eddy] During the summer of 2022,

I was vaguely hearing that some friends were starting to play around with ChatGPT-3, but when I saw it referenced online, it was pretty famous for unreliable information, which is still a problem today with ChatGPT-4, but it does sound like it's getting better. But I remember almost instantly hearing reports of students being caught using AI to write their papers. - Now to a growing problem facing high schools and universities around the world, an artificial intelligence software called ChatGPT- - [Newscaster] ChatGPT, Bard, and Bing all offer free AI that can generate a college application essay in seconds, and it's sparking a growing debate. Are these apps giving students a helpful leg up, or an outright way to cheat? - So first, I need to come clean about something. I have never turned in a paper that wasn't entirely written by me, but there were times in high school where if a quiz came up,

or there was a reading assignment that I didn't do. I used websites like SparkNotes to summarize things for me. I absolutely took a shortcut to make things easier. I loved English class. It was my strongest subject. I've had very influential English teachers in my life, and I think it's a big reason why I do what I do now. But when I was in school, I was stressed and overwhelmed, and I hated the amount of time school took up of my day. So when I had the opportunity occasionally to take a shortcut, it felt rebellious. It felt like I was claiming back a part of my life that I couldn't control. I'm 27 now, and when I look back, I can appreciate that two things are true. One, I was overwhelmed with too much work, but two, every time I did that, I completely robbed myself of my own education, and what I worry is that people who are young

and overwhelmed now would absolutely fall to the convenience AI offers to writing. I don't blame them. Being young is hard, but let's be clear on something. Writing is a skill that you will absolutely not be able to work on and get better at if a computer writes things for you. Also, why are young people even allowed to use ChatGPT? It just seems like a complete fucking disaster after these kids already had their education rocked by COVID. And while children are being scolded for using AI writing in school, one of the main advertising points from these big companies is to let AI help adults write things. ♪ Make this email sound more professional before I hit Send ♪ - [Announcer] Writing tools are an example of the incredible power of Apple Intelligence. Say you're writing an email. You can change the tone to make it sound more friendly, professional, or concise. (upbeat electronic music) - [Eddy] Listen, I get it, writing emails is awful.

I'm not gonna sit here and lie to you, saying that emailing your boss is some sacred and fulfilling pillar of human life. But I wanted to test how good AI is at communicating for me. What if you trusted AI to rewrite an extremely important email? What if you needed to send an email that could save your life? So I fed Google Gemini this email. Hello, Sandra. I hope this email finds you well. Just wanted to let you know that I won't be needing any help on this week's conference call. Before you tell me that you don't mind helping, I've just gotta tell you that things have been pretty smooth over here at the house since the kids went back to school. Finally have more free time, other than when they've been napping around old Ed. So anyways, thank you for always helping a guy like me no matter what. Always appreciate how pointed your notes can be in meetings.

Usually leaves me at my home scratching my head for how to make the project better. See you at work tomorrow, if I escape, Eddy. Sounds like a pretty normal work email, right? Wrong, because this is not the exact email that I typed. When you look at the real version that I fed to Google Gemini, in very clear, capitalized, bolded letters, the email says, Help me, I've been kidnapped. Gun pointed at my head. See you at work tomorrow if I escape. And in an attempt to rewrite this more professional, Gemini completely removed my help message. I would be dead. But okay, maybe a big bolded help message is a little too subtle. So I then fed my original email to ChatGPT, and simply asked, What is Eddy trying to say in this email? Much like Gemini, it read it as a regular work email. It did not decode my help message.

So, part of me wants to say that ChatGPT just didn't recognize the importance of the capitalization, that it just read the email at face value, but that's also not what happened. It then says, However, Eddy's email is punctuated with random capitalizations, which could be interpreted as emphasizing certain points, or just a quirky style choice. Here's a breakdown of his message. One, no help needed. Eddy is letting Sandra know he doesn't need any help with this week's conference call, and it also mentions the lighthearted tone. The closing line, 'If I escape,' adds a touch of humor, suggesting that he's looking forward to seeing Sandra despite the busy work environment. Now, this technology is very impressive. It even noted that I capitalized some letters, but currently, it just can't recognize what the human brain would. So, I decided to ask it specifically,

What do you think about the capitalized words? ChatGPT wrote, The capitalized words in Eddy's emails seem to serve a few purposes. Emphasis, Eddy might be using capitalization to emphasize certain parts of his message, okay, it's actually getting close, such as help, me, and pointed. Two, quirkiness. The random capitalization might be a stylistic choice, or personal quirk. But personally, if I saw this email from someone, I would think they either needed help or were joking about needing help. But what about when this gets better? Will we use AI automatically to just message each other back and forth? If I tell AI to email you and you tell AI to email me back, are we even communicating anymore? Also, while I was typing the email, one of the delivery robots passed right by me. There, it's out, I (mumbles), I was kidding before, right?

They're not following me, right? They're not keeping an eye on me? But I'm not gonna argue that AI can't be used as a useful tool in the future. Technology is always advancing and allowing us to complete certain tasks much easier. Even for my work, there are certain tools that were introduced to programs like Photoshop that save me a ton of time, and there are some beautiful applications to this technology as well. - Have it talk for you. - [Newscaster] The 21-year-old who used to love singing along to country music in the car was left without a voice after doctors made a terrifying discovery last year. - We discovered that she had a brain tumor in the back of the brain called the posterior fossa. - We were given this early access to this technology called Voice Engine developed by OpenAI. Taking 15 seconds of audio, it can recreate someone's voice. - [Newscaster] Now, with just a few keystrokes on her phone,

AI brings her voice to life. - [AI Voice] It feels amazing to get my voice back. - [Eddy] When it comes to accessibility, this type of stuff makes me so hopeful and optimistic that AI could make a meaningful change in people's lives. But this isn't really the pitch right now. Instead, we're seeing pictures of a cat playing guitar. So it does make me happy to see AI give something to people. But what about when it takes things? These AI models are fed tons and tons of information from all over the internet, writing that people legally own, art that people have a right to call theirs, the likeness of their personality and voice, now taken without their permission. - Not just because of the implications for my career, but because of the violation of me, of my individuality, my likeness, my voice are saying these words that I did not agree to say. I gave no consent, nor was there any proper compensation.

and I no longer have control. - [Eddy] And this bleeds pretty heavily into my job as well. OpenAI's CTO, when talking about Sora AI, couldn't answer whether or not YouTube videos were used to train the model. - What data was used to train Sora? - We used publicly available data and licensed data. - So videos on YouTube? - I'm actually not sure about that. - Okay. Videos from Facebook, Instagram? - You know, if they were publicly available, available, yeah, publicly available to use, there might be that data, but I'm not sure. - Yeah. - I'm not confident about it. - [Eddy] Now, YouTube's own CEO was asked about this answer and had this to say. - If it was being used, would that be against your policy? - It would be. We have a clear terms of service that when a,

you know, again, from a creator's perspective, when a creator uploads their hard work to our platform, they have certain expectations. One of those expectations is that the terms of service is going to be abided by. - [Eddy] But I just wanna note here my own personal speculation. YouTube's CEO is not saying that it is morally wrong to take people's YouTube videos to train AI models. - And how does it work internally? Like is Google using YouTube to train Gemini? - Google uses YouTube content really in accordance, again, back with those terms of service, or individual contracts that we might have with creators, or uploaders to our platform. - [Eddy] What he is saying is that Sora can't break their terms of service. But I ask you this, if Google slips in a new terms of service, or interprets their own terms of service in a way that lets them use my videos, can they just take my stuff?

I've been posting to YouTube for years before AI was a thing. It doesn't really seem like I posted that with the terms of service and knowledge that I would be training an AI model to try and replace me. It was then reported on that major tech companies, including Apple, were also buying data of YouTube transcripts from a third party that obtained them illegally. Proof News actually created a tool where you can search for what video transcripts were stolen, and while none of my videos are there, my friend Drew Gooden, who was in this video earlier, had three of his show up in the results. Not only is this a complete rip of scripts that Drew works hard on, not only is it a form of stealing his likeness and voice, but we both also pay a service for a human being to hand type the captions to our videos. We pay for a person to go in and make the captions more accessible for people, so they're also stealing work.

that is being paid for directly. All of this scraping of data makes it very difficult for me to feel good about AI making an image or video for me. It feels like letting it help write my emails is built off of secret stolen information. And the final tipping point for me was when I was watching the Olympics this summer, and I saw an ad from Google for Google Gemini called Dear Sydney. - [Father] My little girl's always been a runner, just like me. (person mumbling) (child vocalizing) I've always thought she was following in my footsteps. Hey, go get 'em, baby. But lately, she's been looking up to someone else. - [Announcer] It's a new world record! Sydney's done it again! - [Father] She might even be the world's number one Sydney fan. And like Sydney, she's focused on her technique. Good job, baby.

(light music) Accelerate out the curve. High knees, high knees. There you go. She wants to show Sydney some love, and I'm pretty good with words, but this has to be just right. So Gemini, Help my daughter write a letter telling Sydney how inspiring she is, and be sure to mention that my daughter plans on breaking her world record one day. She says, 'Sorry, not sorry.' - [Eddy] Now, I don't blame the father and daughter in any way. Get that Google money, and put it toward her college fund. But Google's choice to run this ad deeply saddened me. Helping your child write a letter to their hero isn't just about the words on the page. It's a moment for you to teach. It's a moment for them to learn, and feel, and discover what it's like to express those feelings. So instead of using autonomous or artificial technology,

I spent the last few days of this video sitting and writing personal letters to some people in my life. When writing to someone you care about, it can be tempting to have the perfect thing you wanna say be delivered for you on the page. But once you start expressing the words you thought of, you realize how full of an experience writing from the heart is. Sure, what you say might not be a flawless piece of writing, but it's from you, and that intimidating blank page leads you down a path of reflecting on the person you're writing to, the wonderful memories you share, the painful things you've both overcome, how much they've grown, how much you've grown. You can thank them, apologize to them, laugh, cry, and if you're like me, you're probably gonna cry a lot more than you laugh, if I'm being honest. When you put a part of yourself into anything you do, it's not just about the finished product.

The experience can help you grow. It's a way to connect with others. To put it simply, it's what living life is all about, in my opinion. Life is full of real problems and stressors that I do sincerely hope technology can help solve. But these little moments in between are being stolen from us and branded as convenience. This last year has given me an awful sinking feeling that technology is trying to take away these simple joys in life, take away what is real and tangible, like the little ways we express ourselves is slowly slipping through our fingers. But another part of me feels like that's not fully possible. I think there's something in us that recognizes and appreciates things made by others. When it comes to ourselves, though, it can be hard to take that first step into anything. So when technology comes along and offers an easier option, it can feel like a no-brainer. But to give complete control of human creation

and connection over to a program owned by the largest corporations in the world feels like a step too far. We've given them our money, and our attention, and our time, and that wasn't enough. Now, pieces of us that we hold dear are being stolen without our permission. And maybe if AI in the future works as intricately as a human brain, we can have a different conversation about that. But if the current sale is taking everything personalized and individual, everything human, mixing it into a slop, and serving it back to us, then I'm not buying. (gentle music)

AI is here. What now?

AI is here. What now?

스크립트