Biggest Problems For The Future Of AI Art - Glenn Gers

Truth is that most art in some way breaks a rule, does its own thing. That is why we like it because it is original or unexpected. Why do you say that AI is a lie? — Oh, um, almost everything that is promised about AI is not currently true and there is no real indication that it will be true. um the companies that are serving us our technology frequently operate from theory and I guess the best way I can explain this is GPS which is a miracle. On the other hand, if you put into your phone how to take a certain path in a place that you know well, you will often see that the GPS actually is going a bad way. There may be speed bumps they don't know about or they just may be hard to make a turn. And the thing is everyone keeps saying well the GPS will learn that but at least in the past many years the truth is still the GPS does not always know the best way. They know the mathematically most probable way. So therefore the idea that GPS is telling you the right the best way is if you're going to name it it's a lie. It is not true. It is theoretically a good way and statistically it may be but the truth is we don't live in statistics we live in the real world AI is promising first of all it's promising to become conscious whether or not that is true it certainly is not true now so all it really is probability statistics statistically people will like this sentence um that is the exact opposite of what An artist does is make choices. Making art is making choices. There is not a right or wrong art. better or worse art. There are people who have certain values and therefore like types of art. But what the artist must do is say among these many choices I choose this one. Now an algorithm can choose based on whatever values were put into it which are generally popularity. But most art is not guarante is not guided towards exclusively popularity. Most art that is exclusively guided towards popularity is pretty bad. The truth is that most art in some way is why we like it because it is original or unexpected. If it was simply turnurning out the very same thing all the time, we would actually stop calling it art. We would call it, you know, some kind of stimulation. But art conveys some form of creativity, making a new thing out of old pieces. AI currently is not doing that. It has been promised that it will, and maybe it will, but at least at the moment, it's not doing what they said. And if it were to do that, actually become creative and conscious and all these other things which are could be very bad for humanity, but leaving that aside, just artistically, why would you want it to do that for you? That's like saying, I want to be a writer. I'm going to take my idea. I want to write a story about this. AI could write the outline for me and then I could write it. That's the exact equivalent of me saying, "I want to write this story. Would you do the outline for me and come back to me and then I'll write it? " What kind of artist does that? says, "I don't really want to do the art. I just want to have done it. " While every artist in the world says that, the truth is, if you took it away from if you said, "We can feed your desire into a machine and it will come out with a work of art and you put your name on it. " Very few honest artists would say that's the plan I want to do. They want to work the work and make the stuff out of themselves. And while there are tools that are useful, if you want the AI to tell you 47 possibilities for the next scene, sure you can do that. But you've got that in your head. And if it comes from your head, it's more likely to be good art because it's coming from a consistent place. If it comes from the AI, it's coming from a conglomerated version of a lot of other things and then AB tested to what's popular, which means it's not yours. And if you like something that's popular, it'll already be in your head. You don't need the AI. I just honestly feel that very [clears throat] few uses of AI

help you make art. The truth is it helps you do the business. It helps you skip steps, but skipping steps is not the goal. Once again, it would be like if you were playing tennis. He said, "Okay, but I'm really bad at serving. Can I have the machine serve for me? " Well, then yes, but you're not playing tennis. you are doing a part of tennis and you're having a machine do if you really like tennis that feels bad and the odds are even if you were to win playing tennis that way would you be able to take that credit and say that's my win? I don't personally think so. I do recognize that it can be an interesting tool. The only thing I would say is that you do not ask tools to make your choices for you. You don't put your food on the stove and say, "Stove, decide what I want for dinner. " You say, "I'm going to make some chicken. " And then you use the stove to cook it. Cool. That's a tool. Having it decide for you is some other thing than a tool. I don't entirely know what to call it, but whatever it is, I believe it makes for bad art. I could say that if let's suppose ceramics. So the fun of ceramics is having the wheel right there and getting your hands in the clay and there it's wet and it's just mushy. But what if you don't know well if I put the kiln at this, you know, level there's ways that you can ask it questions that could help the art. — Absolutely. — But not maybe do it for you. — Right. But that would be a book. I mean, in that case, the AI is simply the equivalent of a good reference book. You are looking, you're turning to the AI and saying, "Hey, — you've read 50 ceramics books and you're going to tell me the majority of them say this is the temperature to use. " Cool. But you're not going to say, "What should I make, a cup or a plate? " Because that's not a reference book question. That's a question of choice. And that choice only makes sense from you. You could ask it, if I'm trying to sell plates, what color plate is most popular at the store? However, most artists don't just want to make whatever's popular and will sell. They want to make whatever inspired them to like plates in the first place. And there will have been plates that you saw or ate of that that meant something to you. And if the AI says, "Ah, that's not a popular plate. " A good artist would say, "That's the plate that matters to me. That's the plate I want to make, and maybe I'm going to sell less. " — Guess you could say then it's inspiring people that don't have an artistic streak or writerly sort of bent to be able to do things. — Absolutely. If that is a completely legitimate use of a essentially maybe a toy, maybe a tool, but the point is yeah, like if you want to have something make a song for you and you can't be you don't have talent at writing a song. But the point is there's a certain question of like at what point if you don't have a talent or an interest in pursuing a thing why are you doing that thing that I guess that's the question and there may be legitimate points at which you're saying I just want to do a little of it great use AI absolutely but then honestly you're less of an artist and I feel awkward saying that but I would say that you're more of an artist to do a bad job by yourself. That when you're making plates, if your first thousand plates are wobbly and not circular, that is still more fun and better. And in theory, you're going to spend like if you love an art form, if you like to do the thing, you're going to be like, I'm gonna look, this plate is slightly more round than the last plate. I'm winning. It's good, — right? I mean, you could say that, and I know this is not the case, but it seems as if there's an element of magic involved. And I know there's not. It's a culmination of different things that are that are, you know, in the sort of the online and things like that. But there is an element of like this I'm going to put this in and poof, I'm going to get a rabbit out of a hat. Yes. — And I think that part is enjoyable. It's novel to us now. what AI has only been around what three years in terms of certif like I said it's — honestly to me that's a toy a really great toy and I like toys so I'm not saying that in a bad way but I'm saying it's very different from an art medium um and doing art is very different from playing with toys and I'm not saying that playing with toys is bad I'm

saying if you want to be an artist it is only satisfying to do the art. And it's hard because you're never good when you start, but in theory, at least some little part of it will be fun or at least interesting or challenging and that's why you do it, — right? And are there any pros to using AI? Not pros writing, but are there any, you know, pros and cons? Are there any pros to doing to using AI? I personally do not think so. I personally think that un if you were using it as a reference tool, in other words, you can go to the library and look stuff up or you can ask AI to look it up to a certain extent, sure, why not? But the problem is currently at least the way I is being the way AI is being built, it's not giving you good answers. Much of what it says is not true. based in things that are online. There is a great deal, despite the mythology, deal of knowledge and stuff that is not online. And I know this from like doing research on a historical novel I'm writing. I keep running into these art history books that are not online in any form. They're just books on a shelf in a library. If all of those were in the database for AI, I would say great, but currently you're not actually getting what they're promising. It is not sweeping all of knowledge. There are lots of pictures that are not online. movies and music that are not online and AI should be one place you look but you should also look at those other places. — Let's suppose somebody who uh loves Shopan or or Beethoven said you know when the synthesizer came it's going to just the piano is going to be done. — But there's those two can coexist. So — that's once again like if you were looking at AI as an instrument, as a tool that was like the point of a synthesizer was it made noise in a different way than a piano. A piano hit a string. A synthesizer connected some circuits, but they both made a noise. How you make the noise and what you do with the noise, that's the art. The synthesizer is just a tool. To the extent that AI is doing something for you like the synthesizer in making the noise, that's great. But the minute you say make me a melody, you're taking the noise and the creativity and giving it both to the AI. So what I would say is yes to a certain for instance if you wanted to make a movie an animated movie and you can't animate it's too hard. So you say to the AI make an animated movie where and you script it and you give it and it makes the animation to a certain extent that is legit. If you are claiming you wrote it, but if you're saying you animated it, you are lying. You are saying I am you are taking credit for something that something else did. And that thing would be just as much as if you wrote an animation script and sent to an animation studio in another country and said, "Here, I'm going to pay you this amount of money. Animate it. " And then you brought it back. If you say, "I wrote this. I sent it to a factory. They made it. Here it is. " Cool. But people are saying, "I made that. I made it with AI. " That to me is inaccurate because you didn't do the work of the animation. So if you wanted to say, I wrote a script and then I had AI animated. Great. — Okay. Could prompts be a new art form? — Sure. The problem is the control of the mechanism that is giving back the responses to the prompts. In other words, yes, you can ask that But once again that would be like saying to a person what would you do if you were in a empty room and there it was filling with water. Okay. So you say that and that person would run through the possibilities in their mind and they would choose by their values and they would say a thing. That's exactly what they're saying AI does. The point once again is when did it become part of writing to go up to a stranger whose values you don't know and whose values are being operated by a company that won't tell you how they're doing it and saying to them room full of order. What

do you do? And then they give it back and you say and you take what they did to a certain extent you're not writing. — Okay, that's fair. And it does tend to agree with you. — Yeah, that's that is the one thing. Yes. Yes. — It's your personal yes person. I — I do not It's I'm not saying that the technology could not be crafted in such a way as to be much more useful. It still could not be creative, but it could be a great tool. The problem is currently it is not being offered that way. It is being offered in ways that stimulate engagement and keep people coming back on a very consumer level. But we are not being consumers. We're trying to be creators. — Sure. — And that's different. And the AI is not built to help that. It could be and it should be and if it was, that would be a legitimate tool. Currently, it is not. — Right. Well, just as we've kind of rebelled against social media uh at times and putting our lives online in certain ways. Maybe people AI will have its peak and then people rebel against certain parts of it. I don't know. I'm just — I still use GPS. I just don't believe in GPS. I recognize that GPS is making a statistical suggestion and it probably is good. But if I have knowledge, I add my knowledge and I operate. In other words, that's what I'm saying. It's a tool. It has limits. It has strengths. Know that it's a tool and that much of the algorithm is a black box. It is even a black box to the people who make it. And blackbox means they don't know how it's doing what it's doing. That's disturbing. But it is scientifically stated. There's a lot of people working on the algorithms that run AIs that are saying we don't actually know how it's doing that. That's troubling. It's not that it's not a valuable tool, but just like GPS, if all you do is follow the GPS, also if everyone's using the same GPS, they're all going to be on the same street. that street is not going to be the best way to go because there's going to be too many cars on it. Likewise, if everyone's going to a to AI and saying, "What's a good story? " It's going to narrowly continue to narrow down to a single most popular or most valued thing. It's the exact opposite of art, which is I've heard that I want to do something a little different because that's a big part of art. Otherwise, we would have had a work of art that finished art. The whole point of art is that that's not how it works. That there's no right art. There's different art for different times and different moods and different people and what you want is a whole lot of different art that came from different people. So, if AI is the creator, it's eventually just going to be boring. You know, I love some of these documentaries on these savants who do these recreations of these paintings. There's been several of them. And they sometimes they dupe these curators and then they curators buckle down and say, "I'm going to find them. " And in some ways there's an element of that. I [clears throat] don't know. Maybe. But the with that it's still a human being, you know? And you think if you could do a replica of the Mona Lisa, why wouldn't you want to just attempt to do your own version of something else? Why try to do I get it because they want to try to get — and also that it's simply a different skill. I wouldn't say that making a precise copy of a work of art is art. I would say it's a skill. It's a talent. It's a craft. It's a beautiful thing. I'm not saying it's bad to do. I would differentiate that from anything creative. Any form of creativity, even if it's stick figures, is art. But if all you are doing is imitating another thing to me it is I guess art without creativity and I'm not sure that that's art that's it's craft like I said and it's skill and I'm not saying it's not those things and those aren't amazing but it differentiates from the one defining factor of art which is you're making something that didn't exist before even if it's using a lot of other stuff that particular version it's yours in that moment as best you can. So maybe there could be this new uh curriculum is AI and philosophy put together rights and wrong. You know, it'll start its own uh — as a tool in theory. If it was being molded to be really precise and honest, if it was willing to say this is an approximation based on these sources, there's all sorts of ways to make it a phenomenal tool. None of that seems to

be operating right now. And that's too bad because it's a cool technology. There's nothing wrong with it. — Cool. — The problem is not the technology. It is the people making and running the technology, — right? Well, in my opinion, — as the chef says, I don't know what I put in my sauce. It just, you know, — probably does know, but yeah. — Thank you for watching the video all the way to the end. Here is a complimentary question from our book story questions.