Learning Curve

Is AI Boosting Creativity, or Stifling It?

Episode Summary

Proponents of new AI tools say they will unleash human creativity. But some artists worry that the tools will flatten creative output. For this episode, Jeff interviews two prominent artists with different takes on the issue: Angela Ferraiolo, an art professor winning awards for her experimental AI art; and Ted Chiang, a prominent science fiction writer who argues that popular AI tools are antithetical to the creative process. We also debut a companion website for Learning Curve to help you think through where you stand on the issue.

Episode Notes

On this episode, Jeff talks with two prominent artists with different takes on AI. One experimental artist sees her students doing new kinds of work thanks to AI, and a celebrated science-fiction author argues AI is sold as a shortcut that takes away human agency and encourages a form of plagiarism masked by technology.

Episode Transcription

Jeff Young:

Hello, and welcome to Learning Curve, a look at how education is adapting to the rush of generative AI.

I'm Jeff Young, a longtime education journalist.

This week we are tackling a really big question: How is AI impacting creativity and art? Will artificial intelligence unleash a wave of creativity, or will the presence of this technology actually stifle it? To explore these questions, I talked with two prominent artists with different takes on this issue, and honestly, this is an all-star episode, so you are in for a treat.

This episode got its start last month when I was at Sarah Lawrence College, right outside New York City, to give a talk. And, of course, anywhere I go, I try to find interesting people to talk to for this podcast.

Wherever I go, I ask for recommendations for professors who are digging into AI in interesting and nuanced ways, and someone said there was this art professor that I just had to meet.

Angela Ferraiolo:

I'm Angela Fariolo, and I am a professor. There's no ranks at Sarah Lawrence. I'm a professor of visual arts at Sarah Lawrence College, and I'm also - it's called a computational artist, so I use code. All of my artwork is based on code, and I build systems that change over time.

Jeff Young:

Angela lives in the city, so we arranged to meet up at the New York Public Library. We sat in the plaza in front of this iconic library building. You can probably picture this building - it's the one with the lion statues in front. It seemed like a fitting setting, since this is one of the largest holdings of human creative work in the world. It is also one of the most visited, so you'll hear some honking and city noises in the background, giving flavor to the scene. I started by asking her why she was drawn to incorporating AI into her art.

Angela Ferraiolo:

How I got into AI right was through the video game industry, which is a very different kind of AI than prompt-driven what's called generative AI.

Jeff Young:

Yeah, you were making like the some of the first multiplayer, massive multiplayer online games, right, back in 2002

Angela Ferraiolo:

Yeah, exactly. And it's just seems like another lifetime. We thought it was so amazing and space age back then, but like, of course, now it's really surpassed like anything we could have imagined at the time, but yeah, so to me AI means making things that have behaviors that convince humans they're intelligent. It's not so much this system of typing in a prompt and then getting back a picture, like I have a different set, I have a different experience of what AI is, and I have different goals when I sit down to work with AI, so I almost never use prompt-generated AI, although I do, you know, I like, I have used it, of course, and it's a wonderful experience.

Jeff Young:

Angela's films are hard to explain, they're experimental, and they're designed to be pushing boundaries of art, and her efforts have won acclaim, including appearing in the New York Film Festival and in Sea Graph, the big annual conference about computer graphics. One of her recent works involves using AI to bring different ways of seeing familiar scenes in New York City,

Angela Ferraiolo:

So this movie was to build a camera, like a synthetic camera, and to ask it to look at the world the way an AI might see the world, and it took three different AI, or three different spatial regimes. The first was the Cartesian regime, which is how you and I see, and how a regular camera sees, and it just shows you what's out there, what's in the world, and then the second regime was called a spatial or relational regime, and it might be like if you were in a maze, you could imagine this, so think of my video game background, right, if you were in a in a maze that changed, like you would take a step and the maze would reorganize itself, it was that kind of an intelligence, and the third regime was just an agental regime. So the AI builds its intelligence based on where things are in relation to itself.

Jeff Young:

I'll put a link to her work in the show notes. I started by asking her why she was drawn to incorporating AI into her art.

Angela Ferraiolo:

There's, I think, there's something about seeing an intelligence that isn't a human intelligence, and seeing like there's something unnatural about it, and there's something weird about it, and in a way it becomes fascinating over time.

Jeff Young:

It was just almost like creating this form that you don't know what it's going to do next?

Angela Ferraiolo:

Exactly. I mean, you kind of know what it's kind of like. It's kind of like if you have a kid, like you kind of know the range, but you don't really know what's going on all the time, right? I know parents out there will feel sympathy for this.

Jeff Young:

Do you have kids?

Angela Ferraiolo:

Of course, I do. Yeah, some are very tech. Some are very like love technology, and some reject technology. So, it's like it's sort of an interesting situation

Jeff Young:

These days, Angela is encouraging her students at Sarah Lawrence to experiment with AI tools as well, and so I wanted to hear what they are doing with these new tech tools.

Angela Ferraiolo:

You know, it's interesting who turns out for the AI classes, because it's not the art students, it's the students who wanted to be artists and felt because they didn't start like at a very young age they could never be artists, that is mostly who will take an AI class and make their first movie and do their first drawing and make their first graphic novel using AI, so in that sense it's been that tech promise of the great democratizer. Right?

I mean, that is the AI promise. Anyone can make a movie, things like that. That's what I've seen in my classes,

Jeff Young:

And that's happening?

Angela FerraiolO:

Absolutely, yeah, that's absolutely what's going on in my classes, people who feel like, oh, these kids in my dorm, you know, they've been doing music since they were seven or eight, and they're like, so advanced, and I could never write music.

I'm like, sure, you could write music, like, to sit down with an AI for a week, and I'll teach you, like, you, you won't, you won't have the same skill set your, your, your roommate has, but you'll understand, though. You won't feel like it's impossible.

So, I've seen that, you know.

Jeff Young:

And what kind of movies are they making?

Angela Ferraiolo:

I had a movie made about a student's, like, winter vacations with his family, and, like, sort of like a blur between, like, how I remember this vacation, and like how, like, the good parts and the bad parts, like a very, like, strange fantasy review of vacations. And then I had another student from the Dominican Republic take all of these mythologies of the Dominican Republic and make those into AI movies, and then I had another student who was convinced her supermarket in the Bronx was haunted by vampires and created this supermarket that was haunted by vampires as their AI film projects, like a family birthday that you weren't there for that, you've recreated or someone's wedding, like my great great grandparents' wedding. I'm going to recreate this in AI.

Jeff Young:

Do you think this is something the students get? Something I mean, this is like was not possible before. It's fair to say

Angela Ferraiolo:

It was absolutely not possible before, and you know, in a certain sense, AI's removed the risk they might feel like taking a film course and going out, because you could just sit down and write the script, and like cast people, and like, like go to California, but the economics are different, but also that level of risk is different, right? It's a lot of time, and then you've got to work with a lot of people, so it's like a really different way of making a film, and these students, it, these students appreciated those differences.

Jeff Young:

I'm curious, like, what's the stakes? Like, did they, was they getting a lot out of this, you know, as our, because you could say, like, oh, it's so quick, I don't know how their projects, but oh, if it's so quick, do they get the same depth of the, you know, feeling, because …

Angela Ferraiolo:

It’s because the like those obstacles have been removed, yeah. I think they do. I think they go there. It's much more of a what if kind of proposition, though. So I think it's a lot more like improvisational theater might be, where you're, where you're working with a collaborator and they take the story this way and like you kind of build on it, so it's like a back and forth, so I saw that, and I saw people try crazy ideas.

Like, ‘Yes, I can make that science fiction movie now, because I can do it, so in that sense I thought it was great,’ in the sense of like what's lacking is getting people to show up to hold the mic to hold the boom microphone, and like getting dealing, dealing, dealing with like getting the location cleared through student affairs, and like finding the right person, like how do I direct this person in a scene, like all of that is sort of is missing, right, but then there are other skills that are AI skills, which are, you know, just like, as to me, like a different skill set, and that's really what I saw in the class, is like, wow, this is an entirely different skill set that you'll need to make an AI film as successfully as to make a live action film, but we've, like, we've sort of defined those problems, and we can teach to those, and these are new problems that are a little bit trickier to teach to,

Jeff Young:

Of course, not all artists are excited about the influx of generative AI, and one of the most vocal artists who is warning about the downsides of AI for art. Is the science fiction writer Ted Chiang like Angela Ferriolo?

He has his roots in tech. He worked as a technical writer for years, and his short stories often explore issues of technology. His work has won major awards, including four Hugo Awards and four Nebula Awards. Barack Obama has recommended his short story collection Exhalation, and in 2023 Time magazine called Chang perhaps the world's most celebrated living science fiction author. I connected with Ted Chiang last week by Zoom, where he talked to me from his home outside of Seattle.

Ted Chiang:

My name is Ted Chiang, I'm a science fiction short story writer, and more recently I have written some essays on generative AI.

Jeff Young:

One of those essays that he modestly mentioned ran last year in The New Yorker. It was titled Why AI Isn't Going to Make Art.

Ted Chiang:

I didn't pick the title of that essay. If I had, I probably would have titled it Why AI Won't Make Art Easy to Make, because what I'm interested in is not so much the question of whether AI on its own can make art, but whether generative AI can be a useful tool for artists and my very strong sense is that it, it cannot be a useful tool for artists.

Jeff Young:

He notes that the tools are sold as a way to unleash creativity, but he argues that what they're actually doing is quite the opposite.

Ted Chiang:

I guess the way I would frame it is that I would say that we can think of art as a concentrated form of intention, and generative AI is being sold as a tool which produces large outputs on the basis of small inputs, and

Jeff Young:

in other words, if I type a little bit of a prompt, like, hey, make me a, you know, an art piece, a visual art piece, I don't have to say much, and then it makes something big?

Ted Chiang:

Yes, and I mean, even if you, no matter how much you type, the selling point of the tool is that it is producing it is generating more output than you are putting into it, and my claim is that that will inevitably mean a dilution of your intention, and so any tool that works that way, whatever the technological underpinning is. Any tool that works that way will not be a useful tool for artists. You need artists need a tool, a tool which basically you have large inputs in order to create large outputs.

Jeff Young:

What's an example?

Ted Chiang:

So, I would say that Photoshop is a tool. Conventional Photoshop is a good tool for visual artists, because the user interface for Photoshop is extremely complicated. It takes years to master Photoshop. You have very fine control over the outputs. Any tool that is good for an artist is something that will have, like, a very large user interface that will enable very fine control over the outputs.

Generative AI is being sold as something which you know you only need to, you know, provide smaller inputs. It's being sold as a way to, you know, reduce the amount of effort you need, and that will inevitably mean it reduces the degree of control that you have over the output, and that I think runs counter to the art making process.

Jeff Young:

The other day, I had an experience with AI that I think illustrates his point. I tried out the new version of Anthropic's Claude, the one called Claude Fable, which is so powerful it was briefly blocked by the US government. To see what it can do, I prompted Claude to make a web-based video game, and for fun, I asked it to recreate a game that I made when I was a kid, first learning to program on a Commodore 64 computer. Yep, I know, I'm old. That was a long time ago.

Anyway, making a game back then was kind of an ordeal. I had to learn how to make all the graphics for it, which in Basic I remember were called sprites, and I had to figure out how to make them move when the user hit a key. It was laborious, and I kept having to make changes. As to my planned game, because of all the limitations that I hit, of course, Claude's Fable model spun up a working game that far surpassed anything I made as a kid, and it did it in just a few minutes, based on a short description for me, and the AI tool just made all these design decisions on its own based on my limited prompt. I asked Ted Chiang if this was the kind of shortcut he was worried about with AI creations.

Ted Chiang:

If you are trying to express yourself, if you have a vision in your mind, which you are trying to get out there, the only way that you are going to get it out there is to specify it in great detail, and if you use the

Jeff Young:

AI tools

Ted Chiang:

Well, no, and just in general.

Jeff Young:

I see. Whether it's Photoshop or your clay in your hands?

Ted Chiang:

Yes, or you know, or a pencil or a paintbrush, or or if you are writing code, generative AI, because you know it is taking a relatively, you know, relatively short prompt, and I don't mean that the prompt has to be like super short, but the, you know, the prompt is the selling point is that the prompt is shorter than you doing it yourself, and that is that that reduces the amount of control that you have, so you know, in the, in the context of, say, prose fiction, you know, if you're a writer, you know a lot of times you will sweat over individual word choices,

Jeff Young:

Sure.

Ted Chiang:

Because the individual words matter to you, like they're going to be they're going to be sentences whose very specific phrasing will be essential to the effect that you are trying to create, and you know that is that is what it means to write fiction, but this is true of, you know, any art form. If, if you're a visual artist, there are going to be very subtle effects that you are interested in, in the image, and if you are a musician, there are likewise going to be very subtle effects in the music that you are after, and if you are a video game creator, there are also going to be very subtle effects that you are trying to achieve to create the gameplay experience that you have in mind and conventional tools give you that degree of control over the final product. Generative AI does not

Jeff Young:

Let me ask, I'm curious. Let me ask you this, because this is one of those things, as I talk to so many people for my podcast and my reporting. There's, I feel like, there's so many people that end up hearing, you know, critiques of AI, and they're like, oh, well, there's always something you can do. Maybe it's not the writing of the words like you're describing now, but you know, I know you have worked as a technical writer, you're very steeped in, you understand the technology, how it works.

Like, are there what would you say to someone who's like, 'Oh, is there some way you could use the AI tool to help you plan, or like something in the creative process that isn't the actual maybe writing of the words, but then something that would help you? I'm just, just as just, I'm curious if you have any thoughts of how it could help a creative, but not maybe in the direct making of the art,

Ted Chiang:

I guess I'm not.. I don't know that generative AI is going to be better for artists than, say, you know, the conventional tools that we already have.

Jeff Young:

Sure,

Ted Chiang:

Like, if you are looking for, say, something, a brainstorming tool, I mean, okay, people talk about

Jeff Young:

That, yeah,

Ted Chiang:

Yeah, but the thing is, I mean, we already have brainstorming tools, and they did not require $100 billion to develop, you know, you know, and yeah, they, they, they do the job, so I, so you know, it's not that, you know, you can't use generative AI as a brainstorming tool, but the question is more like, you know, did you need that?

What is it doing that existing brainstorming tools could not do, and you know, if it is appreciably like sort of reducing the amount of effort, then it is reducing your authorship, it's reducing your the amount of control you have over the final product, so that again works against you as an artist.

Jeff Young:

So, what does Ted Chiang think about this idea that AI could democratize. Art and bring more people into the creative process, and what's at stake for the future of creativity. We'll have that right after the break.

Sponsor:

This episode was brought to you by Studiosity. Real learning needs a struggle, but for educators, assessment security just means more detective work, and that's not what students came for. That’s not what the teachers came for.

True learning and authorship isn't found in a probability score or a surveillance log. It's found in the journey.

Studiosity makes assessment security easy by shifting the burden away from every educator and into a proactive institution-wide infrastructure, from Viva-style inputs to clear binary validity reporting. Studiosity is the multi-layered assessment security process, turning institutional policy into practice, just a clear signal to the teacher owned by the student that I did this thinking, de-risk your assessment flow and reduce workload.

Give your teachers back the time to teach, not police, and give your students the agency to show their own growth. Keep the learning real with studios.

Get the walkthrough at studiosity.com/learningcurve. That's studiosity.com/learningcurve.

Jeff Young:

I want to also remind you of my other big project these days. I'm writing the Chronicle of Higher Education's newsletter about AI on campuses called Jagged Intelligence. I track news and give some analysis of the good and the challenging of AI at colleges every week. I have an example of an AI use case, and I report the results of our latest audience poll, and these have been pretty lively lately. People have a lot to say about AI.

If you haven't already, I hope you will head over to chronicle.com and click on the word newsletters to sign up.

Now back to the episode.

Jeff Young:

These two artists are framing the issue of art and creativity differently, and they have different concerns, but I was curious to hear what Ted Chiang thought of a couple of points that Angela Ferriolo made, especially her comment that she sees AI making it possible for more students to do things like make music and movies to express their ideas.

Angela Ferraiolo:

That is the AI promise: anyone can make a movie, things like that. That's what I've seen in my classes.

Jeff Young:

I played Ted a longer version of that quote, the same one you heard earlier in the episode to get his thoughts.

Ted Chiang:

This notion that generative AI somehow democratizes art, I think, is total nonsense. And so the analogy I often use is no one thinks that you can democratize marathon writing by handing out mopeds, you know that that is not what marathon running is. You know, anyone, anyone can run a marathon if they put their, you know, put enough effort into it. You might not run it, run a marathon fast, but you can run a marathon,

Jeff Young:

Right.

Ted Chiang:

And at we need to maintain a distinction between people who actually run a marathon and people who just say, like, I've always wanted to run a marathon.

Jeff Young:

The scooter runs the marathon…

Ted Chiang:

Yes, if you, if you, if you just ride a moped to the finish line, you have not run a marathon, you know, and you'll, it doesn't, it doesn't do us any good to redefine marathon running to include people who ride mopeds to the finish line, and so, in the same way, you know, making art is already, you know, entirely democratized.

Anyone can make art if they try, you know, and you might not make great art, but you know that's not the criteria that it's like, if you are, can you, you know, make art?

Yes, you can make art, as for the idea that, okay, so some people may be having some experience where it's like, oh, they feel like they are making art using generative AI. I would say that what is actually happening is that they are engaging in plagiarism, but they don't know it, and so you'll …

Jeff Young:

Because the models have been trained on the internet and art and books…

Ted Chiang:

And because, again, like, as I've said, you'll, uh. Uh, or as I said in the New Yorker essay, your art requires making many, many choices,

Jeff Young:

Right.

Ted Chiang:

Countless 1000s of choices. And yes, it is hard to make 1000s and 1000s of choices. And if you don't want to make that many choices, plagiarism is one way to avoid having to do that, but you know, most of the time people who engage in plagiarism, they know that they are engaged in plagiarism.

What generative AI offers is a way to engage in plagiarism, but you, it feels like you're not – you aren't even aware that you're engaged in plagiarism.

Jeff Young:

So it's like whitewashing plagiarism the way you're making it sound.

Ted Chiang:

Absolutely is. And there's this programmer named Simon Willison who says that the training AI is money laundering for copyrighted data. And so you know, if you are, you know, you know, if you are the beneficiary of, you know, money laundering, if you're the, you know, you know, things seem, might seem great, it seems like, oh my god, look at how much money we're making from this car wash, because, like, yeah, if you don't know that you know all sorts of crimes were committed, it's like, yeah, you think you think car washes are extraordinarily profitable.

And in the same way, you know, if you're using generative AI to, you know, make something, it's like, oh, this is this, this feels great, but that's because, you know, you are unaware of the crime, the plagiarism that underlies it.

And yes, it feels good to be able to, you know, see things come out of nothing if you are ignorant of the crime that underlies that being unaware of crime, yeah, that feels great, you know, but that doesn't, that doesn't remove the reality that you know that crimes were committed, you know.

Jeff Young:

You're making this distinction, which, for you know, we talk a lot on Learning Curve about cheating and student misuse of AI to put it forward as their own work. But in a way, you're painting a picture where any business person, anybody, whether they're a student or not, using it is kind of doing an act of plagiarism — potentially with these with a generative AI tool like ChatGPT.

Ted Chiang:

Yes,absolutely.

Because legality aside, plagiarism is claiming you did something when you didn't do it. You are claiming authorship of something which you did not create.

In the business setting, you know, being a plagiarist for some report, whether the firm will penalize you for it or not, that's a separate question. But you, by virtue of, you know, offering up a report as your own work, when you didn't write it, you have engaged in plagiarism.

And so, and the fact that, like, well, it's, it's not like no one can trace it, there's no original document that no anyone can identify. That's not the point, the point is you claimed that you did something, and you didn't do it.

And that, so that you know, that's why I'm talking about, you know, plagiarism as a moral failing, as opposed to, you know, separate, completely separate from the legal issues. Or, you know, whatever other, you know, external repercussions there might be, you, you've lied about what you did.

Jeff Young:

I guess you're sort of hinting at it here, but what are the stakes for you?

Why does this matter for society, whether these AI tools are kind of unleashing creativity or suppressing it or creating these moral failures instead?

Ted Chiang:

Well, so you know it is not unleashing creativity, it is normalizing lying.

It is an attempt to sort of change all of our societal norms about what it means to say you did something, what it means to put your name on something. Up until now, we had certain expectations that if you put your name on something, you were the one who made it.

Now you know generative AI is sort of an attempt to dissolve those norms, so that putting your name on something means nothing. It doesn't mean you made it.

And you know, and the generative AI companies will, would have us all say, like, that's okay. It doesn't really matter if you made it or not, you can put your name on anything.

And I think you know that is going to have a corrosive effect on society.

Jeff Young:

In her work, Angela is interested in how AI tools will change art forms like film and television.

Angela Ferraiolo:

What really makes film work?

If you ever, so one of the first things, if you ever teach film, is you show students like the very first movies, which were locked down cameras with no editing in it.

They're unbearable to just one shot. And how it's like watching a play, it's like watching a play, but you don't have the vibe of being in a room with live actors, right? But you're watching a play, and a scene starts like someone walks onto the stage, and someone walks off of the stage, and like, if someone hides, they hide this person in a barrel and put the lid on the barrel, and then there's a reveal. The person comes out of the barrel, and you're just like, "Wow,

Jeff Young:

You have seen these grainy black and white film clips, they're painfully slow to watch, even when objects and people are moving around in the frame, it just feels so static.

Angela Ferraiolo:

From there, we go through montage, and you see how, like, actually cutting film actually creates the story in the film, and you can give,

Jeff Young:

Yeah.

Angela Ferraiolo:

You can think of that great scene at the end of “The Good, The Bad, and The Ugly,” where it's nothing but cuts from one face to another, and it creates the entire tension of the scene, right.

Jeff Young:

In the film clip she mentions from this famous western, the camera cuts from a wide shot showing the two men in cowboy hats in the midst of a pistol duel. Then there's a close-up shot of one man's hand at his side, ready to pull his weapon, and then a close-up of the opponent's hand about to draw his weapon. You get the idea.

The story is being carried forward entirely by the way these shots are framed and ordered

Angela Ferraiolo:

In AI, it's different. It's like I can take this and change it into something else, so like while I'm sitting here, I could change into like a giant like panda bear or a giant like ice cream cone or whatever, I would like to change you to stretch into something.

So that what that theory is about is what happens to film and what happens to us as we watch film as we move from that cut from one idea to another becoming a transformation, because what you really love to watch in AI film is transformation that happens in front of you, right.

So that's just a difference, and there are several differences, and as you watch students working with AI film, you realize they're slightly changing the rules of storytelling here, like they jump around in time, much, much more, like in those jumps are less motivated, and characters can come and go, and story lines can be like very compressed, you know, and things that require dramatization in traditional live-action film, maybe get like just, you know, and then we went, and then we went to Mars, or something like that.

There's a very fast kind of transition, and what's interesting is the actual moment, and how it's changing in front of the camera.

Jeff Young:

Wow, I mean, that seems kind of disorienting, and I've even seen some, you know, little AI videos that are that do this kind of thing, like you say, like somebody changes into a panda, and you're just like it's a little bit disorienting,

Angela Ferraiolo:

It's a little bit disorienting, and it does weird things to your sense of time that, like, the montage doesn't do. So, we'll see what happens with.. I don't know where that's going personally, but I do notice things like the first thing about these new technologies is just what do you notice and how do you feel while you're watching it, but they kind of are disorienting,

Jeff Young:

But there could be good things that come out of this move to morphing?

Angela Ferraiolo:

Yeah, I mean, you know, it's students have a hard time when I tell them photography used to not be considered an art form when photography came into existence.

Jeff Young:

Yeah,

Angela Ferraiolo:

People said that's the original black box.

Actually, if you take our technology and power, it's that camera, right? Is the black box sure people were like the photographer does nothing, photography presses a button, and that's very much what people think about computer art right now, like the computer artist presses a button and out comes the artwork, and I'm like, yeah, that's still how people think, even after you know. Photography after video games, after all of these art forms, which are so much work and so hard won, and had to fight for me, and had to fight for respect. Exactly, here we are, right back again at the bottom of the hill.

Jeff Young:

What was Ted Chang's take on this idea that AI is going to change popular art forms, and maybe how we see the world in the process. Again, he sees things differently.

Ted Chiang:

First of all, I guess the comparison is not to something like the advent of film because advent of film created an entirely new medium, which did not exist before, and you know, so far, generative AI, that is not what it is doing. It is creating stuff in a, you know, is creating, you know, slop in a, an existing medium.

Jeff Young:

I see.

Ted Chiang:

So yeah, generative AI is not, you know, offering some new, like, you know, smell-o-vision, you know.

If someone creates a technology for smell-of-vision, it's like, okay, yes, that's a possible new medium, new artistic medium, but that's not what generative AI is doing, so the comparison is more to something like, you know, the advent of CGI, or visual effects, and, and you know, the example of the morph. I mean, morphing has been around for as long as CGI has been around, I guess. So, yeah, so I see. Do you remember the Michael Jackson video for Black or White? That was, that was..

Jeff Young:

Oh, yeah, do you remember that? Yeah.

Ted Chiang:

So you know, it's not like morphing is new, so, but okay, if we talk about this in terms of like an analogy to things like VFX, VFX offered filmmakers, you know, new tools for their toolbox, and I guess

Jeff Young:

But generative AI is democratizing those effects, let's just say, because you're right, it's not. It is certainly not new to be able to morph, but it's well, now something I could go do, I suppose.

Ted Chiang:

Again, I don't think it's democratizing it, I in some ways, you know, it is, it's making it easy to plagiarize without, you know, feeling like you're plagiarizing, so in terms of whether it's possible for, you know, people to make good art with this, I would say that, or do interesting things with it. It's not impossible, but people are going to have to put a lot of effort into it, because, like I said, making art requires making the old 1000s and 1000s of decisions.

Jeff Young:

Yeah.

Ted Chiang:

So you know, if there are ways to use these tools, where you're like, you spend, like, if you were spending 1000 hours with it, you know, I'm prepared to say that, yeah, you, you are making art, but it's going to, you know, it's going to require 1000s of hours, and that is not the selling point. If you were, if you were spending 1000s of hours, you know, to create a single effect, you are not using the tool the way it is being advertised, because they are, they are being, they're selling this tool as a way of, like, you know, making it, you know, quick and easy, you.

Yeah, if you, if, if it becomes quick and easy, then it, you know, it is. I would say pretty much definitionally not art, because you did not make all the necessary decisions that art requires.

So, if you are using a tool in such a way that you know it basically is requiring as much of your time and effort as if you're doing it the conventional way, but you're getting something different. It's like, okay, that you know, yes, I'm prepared to say that you're making art because you are putting in, you know, the time, you are making countless choices, you in the, in the creation process, but you know, if it, if you were using it in a way to, you know, sort of accelerate your production, then you know you are not engaged in the creation of art.

Jeff Young:

I want to make clear that Angela largely agrees with Ted on the labor involved in making art, and she is not looking for shortcuts.

Angela Ferraiolo:

I think art is, it's really hard to make art, and it's really takes, it's a long process, and it takes a long time to use, like, learn to use a pencil, like everyone has pencils, not everyone's an artist with a pencil, which is the great line, right? I think. Upset David Lynch, or I moved on to Miss, yeah, we'll ask the AI who said that there's a, there's a big idea that it will do everything faster and better, but we've all heard that, like, if you know, you're, if you're over the age of 10, you, you've heard it will be faster and better, and we know it will create its own, like cul-de-sacs, and its own sand traps, and people will battle through them.

Don't think it will be faster and better, will be different. It will create different production obstacles that you have to get through, and it is like, you know I have lived through several kinds of waves of automation, right, and is it is a kind of automation that is something to be concerned about if you're if you have those skills that are now have been automated,

Jeff Young:

And she has plenty of tough questions about how AI companies operate. In fact, she teaches a course at Sarah Lawrence called “Art, Technology, and Power” that critiques the tech industry.

Angela Ferraiolo:

I think it's good to be optimistic.

I think it's good to be realistic, though, because we've had these tremendous decenterings in our recent past, right? We had deindustrialization, digitization, financialization, and it concentrates power among certain institutions, and the consequences are paid by other, other kinds of people, and it's, it's bad, that stuff's bad, and we have no, when you look at the recent past, there's, there's no reassurance that we're going to handle this next wave of automation in an equitable fashion, so all the anxieties justified. I think.

Jeff Young:

Can art make a difference in that?

Angela Ferraiolo:

I mean, art always makes art, art gives people the imagination to make a difference. Art doesn't make a difference. Art inspires the imagination to have it be different, or to maybe find a path that, what, that will be different, right?

Yeah, and without that idea in your mind, you can't make things better, but people make things better. Art doesn't make things better.

Jeff Young:

There's this worry that these generative AI tools kind of flatten things and are not leading to the kind of creativity because the AI is people are happy to just say, like, give me something, and then it's kind of the generic kind of version of what everybody's put into the system over the years, and so is there a worry that you have about less creativity?

Angela Ferraiolo:

So, that's like, always, you know, AI is not that AI is a lot of people see AI as a shortcut, it's not a shortcut, it's an amplifier. So, if I have a student who's going to not write a paper, they would go, like, you know, 40 years ago, they'd go retype something out of an encyclopedia, hand it in as their paper, or, like, you know, 20 years ago, they'd like copy and paste something from the internet, now they put it into an AI tool and generate it, and it's faster, there's even less, they're doing even less to get away, yes, that that student is eternal and lives forever.

Another student equally amplified, right, who goes out to research the paper. This is very true at Sarah Lawrence, by the way.

Jeff Young:

Okay.

Angela Ferraiolo:

Disappears, and you're like, what happened to this person? Can we send out a search party? And, like, they're like lost in the library, like one book led to another book led to another book. They come back to your office, they're like, just can't stop talking, like we get this all the time. AI is a great amplifier for that student, because that student is on the endless search, and AI is a wonderful collaborator for that, you know. So I find it really just amplifies a natural tendency that there is, that there are some people who use it to take the scenic route, and I want to know everything about this, and they've got this great summary of all of our intelligence, right? So I can hold those students to its higher standard. Like, don't tell me you don't know what the Frankfurt School is. You're sitting right in front of Chat GPT, as Chat GPT went to Frankfurt School, was.

Jeff Young:

Yeah

Angela Ferraiolo:

And that student amplifies, and then the other student who was trying to get away with something is amplifying that, like they can get away with something faster, and

Jeff Young:

You're gonna maybe better, but yeah, yeah, like more effective?

Angela Ferraiolo:

More effectively, but look, students who are listening, if you're using AI to write your papers and answer your email, and it takes your job someday, it's on you, man.

Like you put yourself in that position,

Jeff Young:

So you get in what you put in, you get back what you put in.

Angela Ferraiolo:

You get back what you put in, even with AI, and with AI it's just. It's very fast, so it's accelerated, it's amplified, but you, you can ask an AI, like, like, why is David Hume using this term exactly? Like, why does he keep coming back to this, and you will find out, and I think that's wonderful for students,

Jeff Young:

As in many of these episodes of Learning Curve, I find it difficult to sum it all up.

I'm guessing each of you had a very different reaction to what you just heard, depending on your perspective and your experiences.

So, for this episode, we are trying something different. We used AI to spin up an online tool to help you think through where you stand on these questions of AI art and creativity.

I said we made the tool, but actually the idea came from a listener, Tara Baumgarten. She is a longtime learning experience designer and a fan of the podcast. She wrote me an email a couple weeks ago that was actually one of the coolest messages that I've gotten. She not only had this idea for what she called a companion app for Learning Curve, but she had already used Claude code to make the app, and she included a working link in her email.

Naturally, I called her up, and we brainstormed how to share this with all of you.

And I'm happy to say you can try it right now. Just go to learningcurve.fm and click on the page for this episode, or just look in the show notes of the podcast app you're on, and you will see the link, or you can type into your browser bit.ly/learningcurveapp.

It's very simple, but very cool. It gives you prompts to reflect on questions raised in the episode, and you can move some slider bars to signal where you land on the spectrum of nuance that we've talked about. This just takes a couple minutes.

I tried it myself, and I found it really helpful in kind of thinking through these issues. So, please check it out. There is an option to view what Tara calls your stance card when you're done, and if you want to screenshot that, you could post it on social media or just save it for yourself. As in all of our episodes, these are not black and white issues, and I suspect we are all just trying to figure out what we think of all this.

This has been Learning Curve.

On every episode, we tackle big questions at the intersection of AI and education. Please reach out if you have suggestions or ideas for the show. And I am curious, what you think of this reflection tool.

This episode was written and put together by me, Jeff Young, and you can find more about the podcast and my other work@learningcurve.fm Our theme song is by Komiku, with additional music by Blue Dot Sessions.

The episode art was generated by MidJourney, and thanks to the sponsor for this episode, Studiosity.

We'll be back in two weeks with another episode. I'll talk to you then. Until then, check out the reflection tool at Learningcurve.fm