John Donvan

This is Open to Debate. I’m John Donvan. Hi, everybody. I’m a published author. My book, In a Different Key: The Story of Autism, was a New York Times Bestseller, and it was even a finalist for the Pulitzer Prize. And why I’m mentioning it is because I recently learned that its contents, every single word, had been sucked into a massive dataset that’s being used by Meta, the owner of Facebook, to help it train a language-based artificial intelligence whose output is designed to sound as though it’s coming from a real human, a lot like ChatGPT, which I’m sure you’ve heard of. I learned about this from an article in The Atlantic, which said that close to 170,000 books had had their contents scraped for this project. So I checked, and yes, my book was on the list.

So, this is my first unexpected encounter with the burgeoning reality of generative AI, and I’m still working through how I feel about all of this, but I am interested to hear the arguments that will be made by the two sides in this episode who will be debating the relationship between creative humans and machines that learn, and whether the latter should be seen as intruders and oppressors or as partners, and collaborators, and liberators. Here’s the question we are asking, will AI kill the future of creative arts?

Answering yes to that question is the author and film producer Jonathan Taplin. Jonathan has had a storied career. He started as a tour manager for Bob Dylan in the late ’60s. He then produced Martin Scorsese’s first feature film, Mean Streets. From there, he made many more films, but he’s also written books, including Move Fast and Break Things: How Facebook, Google, and Amazon Cornered Culture and Undermined Democracy. His latest book is just out, and it is called The End of Reality: How Four Billionaires Are Selling a Fantasy Future of the Metaverse, Mars, and Crypto. Jonathan, thanks so much for joining us, and welcome to Open to Debate.

Jonathan Taplin

It’s my pleasure.

John Donvan

And answering no to that question, which again is will AI kill the future of creative arts, we have Dr. Rebecca Fiebrink. Rebecca is an artist. She is a professor of creative computing at the University of Arts London. She is the inventor of the Wekinator tool, and that is a machine learning program that connects physical, real-life movements to computer responses, and it allows the user to create all sorts of things, musical instruments or gestural video game controllers being just two examples. Rebecca has had two passions in life. One is music, the other is computer science, and she has put them together. Welcome, Rebecca, to Open to Debate.

Rebecca Fiebrink

Thanks, John. It’s great to be here.

John Donvan

So before we start, I just want to get a sense for each of you, uh, what, what the stakes are in this question, in this debate. What motivates you even to join in the debate? So Jonathan, let me start with you. Without getting into the, the substance of your argument, why does this issue matter to you?

Jonathan Taplin

Well, I spent most of my life with creative artists, and specifically musicians and filmmakers, and in year 2000, uh, Levon Helm, who was the drummer for The Band, got throat cancer, and for many years, he made a very good living off the royalties of The Band’s work, and then in 2000, Napster arrived, and his royalty stream stopped, and he had no money to pay for his cancer care. And so, that spurred me to write Move Fast and Break Things, and, um, you know, from my point of view, the permissionless innovation of big tech has been the problem, and you know, when I get into my opening argument, I’ll, I’ll explain it a little further.

John Donvan

All right. Thanks very much, John. Uh, and, uh, Rebecca, same question to you. Essentially, what, what brings you here today to make this argument?

Rebecca Fiebrink

Yeah. I mean, uh, for me, I can’t think of many things that matter more than human creativity and, and human expression, and I’ve spent 20 years of my career working with musicians, and artists, and game designers. I’m really interested in what, especially AI, but other technologies as well, what tech can offer them, that’s valuable to them. Um, I like building a lot of things with creators and then seeing what they do with it. Um, so I’m excited to, to talk about that, but at the same time, I see a lot of misconceptions out there right now about, you know, what AI is, what it’s for, what it can or can’t do. I want to spend some time today clearing those up, and also, I see some really amazing alternative visions coming from the arts communities for what AI can be and what it should be, that I think offer us, uh, an alternative to the stories we’re being told by big tech.

John Donvan

Thanks, Rebecca, and thank you, John. So let’s move into our first round, and our first round is composed of opening statements, uninterrupted, made by each of you in sequence. You each get up to four minutes to explain your position. Jonathan, you are up first, and you are answering yes to the question, will AI kill the future of creative arts? Here’s your turn to tell us why.

Jonathan Taplin

So, a few minu- months ago, Getty Images sued Stability AI, the maker of the popular image AI image generator, Stable f- Diffusion, for not only ingesting 12 million of Getty’s copyrighted photos, but for also altering and removing all the copyright management information from the images. It seems clearly unlawful, but big tech has successfully skated through legal matters for years, ever since Napster and YouTube were founded, waving the banner of permissionless innovation. I wrote a book in 2016 called Move Fast and Break Things, of how big tech was wrecking the music business and the journalism business, and nothing has really changed. Only the biggest of big tech players will dominate generative AI, because it requires massive amounts of computing power. Copyright appears to be fairly meaningless to Google, Meta, Meta, Microsoft, so while humans, creators, rationally explore and debate this issue, tech corporations are using their work to train the generative machines that ultimately may use, make the artists obsolete.

Marvel wanted to be able to put every screenplay that it had ever paid for into a large learning model, and instead of paying screenwriters, it would simply pay prompt-writers, who would write a three-paragraph prompt of what the new Marvel movie would be, uh, you know, the Hulk meets Tony Stark in Iceland, and then in the second act, uh, you know, someone else comes in, and, and, and you know, fairly simple description, and then in a few hours, they’d have a first draft screenplay, which normally takes them nine months and $750,000. So, you know, Marc Andreessen thinks this is all hunky-dory, because it will make Hollywood more efficient. What it will turn out will be something as banal as your attempt to have ChatGTPT write a Stephen King. And it will be more mediocre formula stuff, because obviously, AI can’t really conceive of original ideas. It just simply remixes what’s already out there, uh, in its large language model.

Uh, there are 2,000 books on Amazon that have ChatGPT as a co-writer. Um, these are mostly formula romance novels or how-to novels, uh, how-to books, and I don’t think they’re adding to the corpus of, of genius. And, uh, you know, I will always come back to the point of having been lucky enough to work with a genius like Bob Dylan, to say that genius counts, and genius is what moves the society forward, and the fact that someone is able to mimic Drake and Weekend, and put up a fake version of their song on Spotify, using AI, does not advance anything. It’s just one more version of people taking stuff from real artists and not compensating them for it.

John Donvan

Thank you, Jonathan. Rebecca, you are answering no in response to the question, will AI kill the future of creative arts, and it is your turn to tell us why you’re answering no.

Rebecca Fiebrink

First of all, I have a PhD in computer science. I am a trained musician. I have worked for 20 years with creators using AI, and, uh, I totally agree with

laughs) what Jonathan said, that AI is not going to be capable of making any kind of great creative work on its own. Um, I think it’s always going to require really substantial human creativity, and we can talk more, more about why I believe that, but this isn’t something I see as really contentious. Um, next, we do have a long history of people claiming that new technology is going to kill art. Um, in 1839 for instance, the painter Paul Delaroche saw an example of an early Daguerreotype and he declared, “From today, painting is dead.”

Um, and of course, painting didn’t die. Arguably, painting became more interesting with the dawn of modern art. Um, photography became an art form in its own right, and today, of course, we have photos, thousands of photos on our mobile phones of our friends and families, and, uh, we’re not worrying about how we’re going to afford a, a portrait painter to capture those images of the people we care about. Um, there’ve been similar fears around 3D animation, music recording, music synthesizers. These fears haven’t materialized largely because people care so much about creative work, um, and instead, the latest technology has driven new art.

Finally, every day, I see examples of how AI is actually benefiting creators, and I want to talk about three types of benefits here. Number one is that AI can help creators work more effectively and more enjoyably. So let’s talk about mix, music mixing and mastering, for instance. This is notoriously difficult and time-consuming, and today already, we have AI tools that allow novice musicians without a big budget to mix and master to a reasonable standard, not an expert standard, but a reasonable one, and this allows them to have better-sounding recordings. AI tools in Adobe’s new Firefly suite, um, allow Photoshop users to do more powerful, more accurate image editing.

Number two, AI can help more people participate in more creative activities. One example from my own work is I’ve worked with, um, music teachers and therapists building instruments for kids with disabilities. Um, I’ve built tools for them to build new instruments from examples of movements that the kids find comfortable and examples of the sounds they want the computer to make as a response to those movements, and this is super easy, and it’s really fun.

Number three, AI can enable completely new kinds of creative practices and creative works, and I’m not talking at all about AI creating art on its own. Um, the professional artists I love are using AI to work with media and data in totally new ways. So for instance, Sofia Crespo has done amazing work training her own AI systems on nature photos, and she produces these stunning, gorgeous images of imaginary creatures. Um, Stephanie Dinkins is a New York-based artist who, uh, in one recent work, created what she called a multigenerational memoir of a black American family, told from the perspective of an AI of evolving intellect. She trained her own machine learning system on oral histories from three generations of women in her family, and this lives in a sculpture that you can walk up to in an art gallery and have a conversation with about these histories.

I have dozens more examples like this. Um, none of this suggests to me that AI is killing the creative arts. Yes, there are risks and disruption. We should be absolutely critical of big tech. We should be talking about how to minimize the likely harms here. But this conversation should also recognize all the potential benefits that are out there. We need to minimize the harms and make sure the benefits are widely realized, because what I want is a future where we’re supporting human creativity, human joy, connection, expression, and I think that AI can have a role to play in that.

John Donvan

Thank you very much, Rebecca. Thanks to both of you. Now we know where you stand and why. We’re going to take a quick break, and when we come back, we’ll get into some discussion on our question this week, will AI kill the future of creative arts? We’ll be right back with more Open to Debate. Welcome back to Open to Debate. I’m John Donvan. We are debating this question, will AI kill the future of creative arts?

We’ve heard opening statements from Jonathan Taplin and Rebecca Fiebrink, Jonathan making the case that as he said it, genius counts, and AI is not genius, but AI will crowd out real human genius, and in doing so, uh, threatens the careers of artists and the future of artistic production, um, in a sense by creating a s- a sort of, uh, self-referential, inbred world of art where these machines are just reproducing the same ideas over and over. He’s also very, very critical of the behavior of businesses that are involved in, uh, creating generative AI, again, uh, crowding out humans and using their material, uh, certainly in terms of copyright violation he is saying, using their material without compensating them and crowding them out.

So Rebecca Fiebrink pushes back very hard on this, uh, talking about her own career as an artist herself, and seeing and using and working with artists who are using AI. She points out that there is a long history of fear of new technology threatening the work of art, the example of photography. Photography, she said, did not kill painting. In fact, it pushed painting in new directions and created photography as its own new art form, and she talks about the benefits to artists of, uh, AI, essentially giving people tools to get over some of the tedium of the production of art, enabling the disabled to participate, and actually creating a new kind of art in itself. So we see the two dividing lines there, and Jonathan, what I want to ask you to do is to respond to Rebecca’s point, that in fact, all kinds of new art can be created thanks to AI.

Jonathan Taplin

Well, two years ago in the fall of 2021, Marc Andreessen opened a, uh, a portal called OpenSea, to sell NFTs, and he said, “This is going to be a machine-generated art. This is going to be incredibly valuable,” and so, in the fall of 2021, thousands, hundreds of thousands of people bought Bored Apes, which were machine-generated art, for around $50,000 a piece.

John Donvan

For people who can’t see you, I want to… I just want to point out that you did a big air-quote around the word “art.”

Jonathan Taplin

(laughs) Right. Uh, The Wall Street Journal reported last week that if you were lucky, you could sell a Bored Ape for 50 cents. You paid 50,000. Now you can sell it for 50 cents. I am still waiting for the machine-generated art to do something that my friend, Ed Ruscha, does every day. Uh, I haven’t seen it yet. And as to the notion that we’ve always had new technologies, look, I’m a collector of photography. I love photography. But it still involves a human making human decisions. What disturbs me is a guy came up to me last week at a, a conference in New York, said, “Look, I want to get into the music education business. I want to teach kids how to use Google’s AI music generator to make music.” Now, essentially what this means is, “I want to teach kids to be able to write a prompt to make a piece of music,” and so you could ask the Google AI generator, “Write me a Bob Dylan song against the War in Vietnam in 1965.” And it could do that, and the output would be a bad kind of m- mimic of Bob Dylan, filled with banal lyrics.

Rebecca Fiebrink

I certainly wouldn’t disagree that typing in a prompt to make music isn’t being a d- a musician. I, I think we agree on that. I still think there’s creative value in, um, you know, any number of these Google tools, for instance. I, my favorites that they’ve come out with, there’s been a, you know, a little th- Bach Chorales sort of creator, where you can click on the screen and put some notes in, and it gives you something that sounds kind of like a Bach Chorale. And there’s another one that’s called Blob Opera, where you move your mouse around the screen and you get these sort of operatic voices that pop out in, in nice-sounding harmonies.

These are absolutely not making good art. They’re not making good music. They are going to make pastiche at the best of times. That said, they’re a fun experience, and I think there is a space of fun experiences for kids, for novices, that gives people a very different experience from, you know, sitting back and watching TV or being passive consumers. It, it is something that brings people joy, even if I wouldn’t call it serious art in any sense.

But I think to go further than that, that is a very tiny corner of where I see both the creatively interesting and economically promising uses of AI. Um, I think a lot of the important tools that are going to make a real difference for most professional artists and musicians use AI and much simpler, smaller ways, where it doesn’t feel like you’re interacting with some other intelligence. I mentioned Adobe’s Firefly tools, for instance. Um, you know, there’s a lot of power in allowing someone to make edits faster. There are tools that people are making, not trained on, you know, lots of copyrighted music and p- creating the music for you, but that will take your music, do small edits, allow you to share it quickly-

John Donvan

And so Jonathan, what’s your response to what Rebecca’s saying there?

Jonathan Taplin

So, first off, Rebecca mentioned that, gee, AI could do, uh, could mix your record for you. This is nonsense. You know, I have spent-

Rebecca Fiebrink

I didn’t say it could mix it as well as an expert. I think, I said it could mix it to an acceptable standard.

John Donvan

So Rebecca, can you, can you, can you clarify, Rebecca, what it is you’re saying so that Jonathan can respond appropriately?

Rebecca Fiebrink

I would say that AI can mix your record for you better than you could do yourself as someone who knows nothing about music editing.

John Donvan

Okay, Jonathan.

Jonathan Taplin

So mixing is the most subjective part of record-making, and as anybody who has actually been spent as much time as I have in a recording studio knows, that the idea that a musician or a producer would let AI do the mix is just folly. It’s not going to happen. Even a beginning musician has his idea that the bass is too loud, or the, I’m sorry, but those voices, I can’t hear them correctly, or put some echo on this. This is not something that you enter into the music business with no knowledge. It’s, it’s-

Rebecca Fiebrink

No, I’m, I’m talking about the standard for novices being higher. I’m not talking about replacing experts.

Jonathan Taplin

What Hollywood wants to do is to be able to pay an actor $750 for one time, to come into a studio and get their body 360-degree scanned, and then the studio would own that actor’s body for all uses in perpetuity, for no more money. Now, obviously, this originally is planned, “How do I create a bank of extras that I can put in the back of every scene? And I can put different clothes on them. I can put different, different hairstyles on them.” But anyone who understands movies, and has been on as many sets as I have, know that the only way to get into the Screen Actor’s Guild is to be an extra who a director gives one line to, and if you get one line, you can get into the Screen Actor’s Guild, and then you can become a real actor. So essentially, what this is doing is eliminating the entryway into the business.

John Donvan

Uh, so Rebecca, I think Jonathan’s foreseeing a future in which, in which working artists and, and, and, and, uh, uh, people who can make a living from their art as, as actors, or graphic artists, et cetera, will, because of the presence of artificial intelligence that can do some of these tasks pretty well, will make th- make those careers extinct, fundamentally, and the, and, and, I think he thinks that process is… He’s arguing that process has already begun, and that that’s a way in which AI is a threat to the future of the creative arts, by taking so many creative artists out of the equation. So what’s your response to that?

Rebecca Fiebrink

I would say, you know, as I mentioned before, I think the vast majority of creative jobs are not things that can be replaced with AI, and I, and you know, I would put music mixing and mastering absolutely there. We are not going to replace human experts. That’s just not going to happen. Um, I think, you know, to the extent that there are activities which way, may well include being an extra in Hollywood, to the extent that, to which there are activities that are under threat, um, yeah, we should be talking about how to mitigate those risks, and I think stronger labor protections should be on that list. I think we should be talking about how do we ensure that content creators can opt out of their content being used by AI systems the way that, that you were talking about, John, or we should be talking about ways to compensate people who opt in to having their, their content trained on by an AI system. And I think we’re talking about in the EU right now, there’s legislation upcoming. The, um, EU AI Act is being drafted. There’s a lot of content there that, that people are discussing about how to make it transparent when content has been made using AI, so that consumers are aware of this. There’s, there’s a lot to do.

John Donvan

I, I think I hear you saying that many of the th- the concerns and threats that Jonathan is laying out can be mitigated if thought through, that there, that there’s not an inevitability to them.

Jonathan Taplin

You know, quite frankly, I think Rebecca and I agree a lot more than we disagree on a lot of these things. In other words, my main s- concern is the immiseration of creative artists, and certainly when Stable Diffusion steals all of its phot- photographs of working photographers, and then refuses to pay them anything for them, um, and uses the phony notion of fair use to, uh, say, “Well, we’re just training this model.” I mean, Rebecca and I both give lectures, and we can show some video under the fair use thing to our students, because we’re educating them,

laughs) right? But the idea that Stable Diffusion is educating its AI, which is then going to go off and produce photographs for ad agencies and everything else, without paying photographers, which are the basis of the thing, is nonsense.

John Donvan

Let me pivot to something I’d like us to listen to. Um, you, you refer to this, Jonathan, that, uh, uh, a TikTok creator, uh, released back in April of 2023, um, a, a, a p- a piece of music, um, that it appears was composed by a human being, but the voice, it was an AI-generated version of the singer, Drake. Drake w- had

laughs) no participation in this project, it is believed. Let’s just listen to what this sounded like, and then let’s, let’s talk about its meaning.

singing). Okay, so that’s, that’s just we wanted to give a little bit of a taste of it, so that you could hear that the voice really sounds like Drake, and to make the point that is not Drake. That is an AI-generated, or, or manipulated voice. Jonathan, what does that, what does that signify to you, what we just heard?

Jonathan Taplin

Yeah, obviously, to get a, an AI that sounds like Drake, you had to put Drake’s voice into the

laughs) la- language model. Uh, then you wrote, “Give me something that sounds like Drake,” and I can do that on Google’s music model right now. I can ask for it to give me a Bob Dylan song, and it has a decent approximation of Bob Dylan. It’s not perfect, but it’s decent. So, look, my worry is this. I, I agree that there are tools that AI might be able to u- do, and certainly in the visual effects business in Hollywood, AI has a lot of, uh, use, in, in the same way that AI is probably better at reading a mammogram than a human. But what I w- do not want is to think about the way that AI takes the jobs of other people and, and makes them essentially have no purpose.

Sam Altman said that he expects the marginal cost of intelligence to fall close to zero within 10 years. He said the earning power of many, many workers would be drastically reduced in that scenario. It would result and the transfer of wealth from labor to the owners of capital and the owners of the AI so dramatic that it could be remedied only by a massive countervailing redistribution known as universal basic income. Now, I am very aware that AI is going to eliminate a lot of jobs of white collar workers in the world, but I don’t want it to eliminate the artists. The, I want them to be the last people standing in this onslaught.

John Donvan

Okay, so Rebecca, Jonathan said a lot just then, bouncing off of that piece of music that we heard, so I want to go back to that piece of music and just ask you what it represents to you.

Rebecca Fiebrink

I think this is an unethical and, and possibly illegal use of somebody else’s content. Um-

John Donvan

You don’t think that it essentially illustrates that AI will crush future artistic creativity?

Rebecca Fiebrink

No, I don’t think so. Um, I think we need to have clear legal protections around people doing that, and again, we, we’ve got some of those already. There’s going to be more coming out. I think there’s potentially something interesting there for artists who do want to release, say, models of their voice, Grimes has done this, Holly Herndon has done this, to say, “You know what? Here. If you’re a fan or a content creator, you can use my voice, and we’ll split the royalties on it.” But, um, you know, I, I don’t think that that voice cloning deep fake technology is in itself, um, going to give us the kind of really satisfying creative innovation, either as creators or as listeners.

John Donvan

Um, we’re also hearing the argument made by Jonathan that in the, in the broad, broad sphere of things, bringing in machine learning, which, which is not human, and you’ve conceded is not genius, ultimately will lead to an overall kind of just degradation of the quality of art, that on the, on the very, very big picture, if too much is handed off to the machines that are, that don’t represent genius, that the overall corpus of the world’s art is just going to, uh, it’s, it’s going to decay. I’d like to ask you, uh, first Jonathan, just to… You know, what, what, what are we talking about when we’re talking about creativity? And then let Rebecca respond to that point, about whether the, the world’s art is just going to decay in its quality the more the machines take part in it.

Jonathan Taplin

Well, look. I was lucky enough to be around in the 1960s, when there was an explosion of creativity, both in the music, and then in the early ’70s, in the film business. And it came out of individual artists making really individually important decisions. But today, the problem is that 80 to 90% of popular culture is formulaic. What I think we’re going to do is just get a lot more formulaic stuff. Rebecca and I both agree that this is never going to create anything that changes the world. This is never going to create Like a Rolling Stone. It’s never going to create, create Demoiselles d’Avignon. It, it’s, it’s just not going to happen. So what we’re going to get is a huge amount of mediocre content, because AI is able to churn this stuff out.

John Donvan

But I, I think I hear Rebecca saying her perception of AI making art is, generally speaking, in collaboration with human beings, not just put the box off in the corner and ask it to paint a picture, that the possibility for collaboration could actually cause an explosion of art. Am I getting that right?

Rebecca Fiebrink

Absolutely. You know, there’s been research on human creative processes in all sorts of domains, looking at, you know, what helps people be creative. One of the most important things is how many options did they explore? How much did they play around? How many different things did they try? And actually, a lot of what AI is offering to creators is the ability to make that kind of experimentation just a little bit faster, just a little bit easier, and that’s not the headline-making stuff about AI composing songs. It’s the little stuff around, hey, I can do this thing in Photoshop easier, or hey, I load my music tracks into my, my editor, and it automatically recognizes what the instruments are. These are not sexy things, but I think as a whole, they’re going to help people.

Jonathan Taplin

So this is Marc Andreessen’s efficiency argument. He says AI will transform Hollywood because it will make it more efficient, but you can go down to the Bob Dylan museum in Tulsa, Oklahoma, and you can see 10 handwritten drafts of Like a Rolling Stone. Great art is not efficient. Tolstoy was never worried about efficiency.

John Donvan

Okay, we’re going to take a break right now and continue our discussion when we return, and we will also be bringing in some more voices to move the conversation even further along. The question we are debating, will AI kill the future of creative arts? This is Open to Debate. I’m John Donvan, and we’ll be right back. Welcome back to Open to Debate. I’m John Donvan, and I am joined by Jonathan Taplin and Rebecca Fiebrink to debate this question, will AI kill the future of creative arts? So we would like to bring in some other voices now, members of a small audience that we have here, some journalists that we have invited to listen into the debate as, uh, they are folks who think about and write about these issues. So first up, we have Chloe Stead. Chloe is an art critic, and is currently the assistant editor at frieze magazine. Uh, Chloe, thanks so much for joining us, and come on in with your question, please.

Chloe Stead

Thank you for having me. Um, I have a question for Jonathan. In a recent roundtable, um, in Spike Art Magazine, on AI and labor, um, the artist Sebastian Schmieg, um, said that, “Whenever I realize that what I have c- I have consumed has been produced by Ch- ChatGPT, I feel this instant deep boredom, which I think could push people, writers for example, to really find the next level.” Can you relate to this sentiment at all?

Jonathan Taplin

So, he, he says he feels boredom when he reads ChatGPT?

Chloe Stead

The, yeah, boredom, and that that would actually mean that kind of writers or creators-

Jonathan Taplin

Oh, okay, so it, it-

Chloe Stead

… in general can-

Jonathan Taplin

So it inspires him to prove that the human is more capable than the machine?

Chloe Stead

Yeah.

Jonathan Taplin

I mean, it seems to me that this is the basic argument. This is the whole transhumanist notion, you know? That somehow, we’re all going to merge with machines in the next 10 to 20 years, and look, I understand that Peter Thiel would like to take humans out of the equation and let the machines do everything, but we’ve had, you know, 80,000 years of culture, and humans have constantly advanced things. And so I’m all in with your friend, the artist, who said that it inspires him to do more, because what I’m really worried about is why are they taking, why are these machines taking all the artist’s work without permission and not compensating them?

John Donvan

But Jonathan, I, uh, I, if I understand Chloe’s question, she, I think she’s saying that there might be a net benefit to the creation of art by

laughs) by in a sense inspiring humans wanting to compete with the AI, and prove that they’re better and deeper, and so-

Jonathan Taplin

I, I, I, I, I understand that, John, but that’s shooting fish in a barrel. It’s not hard to be better than AI, you know? It, it’s really almost any artist can do it. My grandchild can do it. So, I mean, uh, this is not a, a tough competition, and if an artist needs the notion that he’s competing with AI in order to get up in the morning, I’m sorry, he needs a better

laughs) job.

John Donvan

Chloe, uh, thank you so much for joining us. We really, really appreciate it. Um, I now want to bring in, uh, Jason Scott, and Jason is an archivist at the Internet Archive. Um, and the Internet Archive, if you don’t know, provides, at no charge, access to a lot of digitized material, including books, and movies, and sound effects. So Jason, please, uh, come on in with your question.

Jason Scott

Hello, both of you. It’s been pretty fun to listen to the, uh, discussions back and forth. There were two main points, or I guess topics that I just wanted to query, I guess both of you to weigh in on. Uh, the first one is it feels like a lot of r- initial reaction by a lot of creatives is to create what I would call a look right. “You looked at my thing, my story as a castle. Now you think that whenever you tell a certain kind of story, it’s a castle.” And, uh, saying that that’s plagiarism. Uh, it s- it seems to me like that really opens up a door where only the big content creators are going to enforce that.

And the other one was, is the main issue unauthorized scanning of this material, that is to say if everybody has shaken hands and everybody knows what’s happening, is that, uh, a case of that’s the natural progression of the world, or is that a problem as well? Because for example, Getty announced their very own AI, trained on their own work, which they’ve supposedly paid everybody for, but which I’m not sure that the original photographers ever knew was going to be used, uh, until last July.

John Donvan

Jason, can you, can you remind me your first question again? I’ve got the second one clear.

Jason Scott

Sure. Uh, the, the thought, the thought of a look right, or a, a listen right, about, you know, the natural response by artists to say, “Well, even looking or being inspired by, or taking from what I do represents a violation, and not fair use.”

John Donvan

Okay, thanks. Jonathan, you wanted to respond to the first question, so why don’t you take that?

Jonathan Taplin

I think both Rebecca and I agree that artists should be compensated. It’s not like we’re trying to stop Stable Diffusion from ingesting pictures, and allowing artists to make new pictures. I loved her idea about some artists making animals that never existed

laughs) you know? Okay, but what we are saying is that the content that it’s trained on should be compensated in some way, and, and that’s, that’s something that even Sam Altman is open to. In other words, there are real discussions going on right now, between the big AI companies and the big arts organizations, as to how artists get compensated. And I feel it’ll get worked out in the next two years.

John Donvan

Rebecca, did you want to add anything?

Rebecca Fiebrink

Yeah. Uh, I’ll add, you know, another really positive development that I’ve seen recently, um, f- coming from the US, is that, is there’s a new project at Stanford called the AI, uh, Transparency Index, I believed it’s called. And so this is something that people are trying to put in place, even before there might be regulation or legal structures, to say, “Hey, if you’re a big company putting out AI models for other people to use, you should be transparent about how was it trained, what was the data, what are the risks that… You know, what’s the risk assessment that you’ve done? What are the limitations you’ve identified?” And currently, none of the big companies are doing that well.

On the totally other side of things, I think it also makes sense to think about more grassroots efforts that artists themselves can take. These a new sh- tool that was just launched recently called Nightshade, which I think is kind of wonderful, which allows you, if you’re going to put images of your own creation on the internet, you can change the pixels in a particular way so that if somebody trains an AI on it, it messes with their model, right? There’s all sorts of models for this kind of pushing back that, that I think are really healthy and necessary.

John Donvan

Jason, thanks very much for your question, for your-

Jason Scott

Thank you.

John Donvan

… two questions. Um, now I’d like to bring in, uh, Eileen Kinsella. Eileen is the senior market editor at Artnet. Uh, Eileen, thanks so much for joining us, and uh, please let us know what your question is.

Eileen Kinsella

Hi, thanks for having me, and um, Jonathan, I love that you mentioned Ed Ruscha, one of my favorite artists, and, um, my question, um, because I cover the art market continuously, I wanted k- to kind of drill down on the question of, um, killing crea- creativity as it relates to the fine art market. Um, and the, the notion of killing creativity because AI isn’t creative or killing creativity because it is, um, d- is destroying the economic basis for creativity. Like if you’re an illustrator, maybe it helps you make certain illustrations, but maybe it’s also like devaluing your career to the point where it morphs or, or disappears. Um, I wanted to know how much those two factors are present as you always consider this question.

Jonathan Taplin

So, I would say that the second issue is the main one. In other words, uh, I agree with Rebecca that there are tools that artists can use to m- to enhance things. But, if it devalues the thing, and, and, you know, I have a friend who runs an ad agency and says that I c- he could see, five to seven years out in the future, where he wouldn’t need all the creative directors that

laughs) he has employed, because many of the prompts [inaudible

:42] “I need an ad for Procter & Gamble’s toothpaste,” uh, you know, uh, the copywriting could be done by the thing. Uh, I mean, certainly, there will be a, a high-level person will come in and edit it and everything, but many of these jobs, people writing press releases, all of these kinds of creative jobs, will be eliminated. And, and so, ul- ultimately, everybody’s job and value will go lower, you know? And, and so, uh, that’s more my concern than your, your first issue.

John Donvan

Jonathan, I’m, I’m, I’m curious, also, and actually, maybe it’s more of a question for Rebecca, that these sort of, uh, less, um, esoteric types of, of creative production, like creating websites and writing ad copy, et cetera, that those are apprenticeship periods during which artists sort of learn… They put in their 10,000 hours of learning the bones of what th- of the work they have to do. And, you know, Andy Warhol, you, you know, was, was doing a- ads for, for M-, for, for Madison Avenue before he started doing, really, his own art. And, my question, Rebecca, is is there a danger that the elimination of these jobs, because they can be automated through AI, something that will sort of, uh, eliminate that kind of apprenticeship period and challenge the, uh, ability of artists to develop their skills, their identities, their voices, even if it’s churning out ad copy for soap?

Rebecca Fiebrink

Yeah, I would say that is a danger for some of the things that AI is likely to be competitive with, you know, an intern. Um, I would also say, of course, that the economics around people starting out in the arts have a lot of problems already, that aren’t due to AI, and you know, as someone who works at an art university, I see a lot of problems f- you know, with students coming from backgrounds where they have to get a full-time, paid job right out of university, and you know, a lot of sort of traditional apprenticeship paths that might lead someone to, you know, becoming a full artist down the road, are, are already not tenable. So I think there are issues there. There are problems. AI is potentially exacerbating that.

Um, but o- one of, one of the hesitations that I have around discussions like this is that it’s very easy to blame AI, right? And I don’t like seeing AI as a smokescreen for what are essentially, you know, power grabs by big companies. I don’t l- I, I don’t think it’s helpful to propagate the sort of stories around AI that are coming from certain areas of the tech sector as, you know, being a technological inevitability, that all these jobs are going to be lost, and we have to, you know, turn to universal basic income, and we have to… You know, society’s going to have to do this if we want, quote-unquote, “progress,” right? I think we should be challenging those stories, and saying, you know what? These companies, these small numbers of individuals, they don’t have a monopoly on what AI is going to be built, what should be built, whether people are going to accept that, um, and I think we, we need to look at things from that bigger perspective as well.

John Donvan

Okay. Thanks, Rebecca. Now, Ei- Eileen, I want to thank you very much for your question, and we are heading down the home stretch of this debate now, and moving into our third round. And in that round, each of our two debaters makes a closing statement to tell us why they are arguing yes or no in answer to our question, uh will AI kill the future of creative arts? Jonathan, as you went first in the opening round, you’ll go first again now. You have two minutes to tell us one more time why you are answering yes to that question.

Jonathan Taplin

So, music existed long before language. The earliest signs of music were found in Slovenia and France, in the form of 80,000-year-old flutes made by Homo Sapiens out of animal bones. Margaret Mead surmised that the first sign of civilization was a healed femur. In the animal kingdom, a broken leg was a death sentence, but the healed femur proved that someone had carried their wounded human friend to safety, and took care of them until it healed. Since they didn’t have language, maybe they played the flute to ease their friend’s anxiety. Our world is 4.5 billion years old, and yet what we call culture didn’t emerge until recently. In fact, genetic studies now reveal that our ancestors might have dwindled to as few as 10,000 individuals, some say even fewer, making humans as endangered 80,000 years ago as the rhinoceros is today. And then our numbers began to grow and human culture began to flourish, and our species, having come perilously close to extinction, reached a point of no return.

Along that continuum, from the flute-playing caveman, to the paintings of, uh, Lascaux, to, right up to Maya Angelou and Bob Dylan, the artists have always made, uh, advancing human genius the key thing. So our task is to just ask ourselves whether we want to continue down the road of techno-determinism or we want to join a resistance movement against machine culture. I believe that that movement, every day, I see it on the picket lines in Hollywood studios. I hear that resistance in the human-made tunes and handmade songs of Rihanna and Gibbons, and I see that resistance in the 10,000 signatures on the Author’s Guild letter to the AI barrens, asking to be paid when their AI systems train on their books.

John Donvan

Thank you very much, Jonathan, and Rebecca, you get the last word with your closing statement. One more time, your answer to the question of whether AI will kill the future of creative art is no. One more time, please, to tell us why.

Rebecca Fiebrink

All right. First, I want to thank you and Jonathan for a, a really fantastic conversation today. Um, we’ve discussed several reasons I do not believe AI is going to fill, kill the future of the creative arts. Um, I think we all agree it’s just not powerful or intelligent enough to replace humans doing truly creative work at any sort of mass scale. Um, we’ve talked about the fact that there’ve been, you know, disruptions before in the arts, from photography, to music recording, synthesizers. Um, the landscape of jobs in creative work has changed continually over the past couple centuries, but people’s drive to keep being creative has pushed us through. Um, I’ve talked a bit about how AI is already benefiting people. I talked about how AI tools can help creators work more efficiently, um, making media editing faster for instance, and I see that as something really important to doing people, creative work people find satisfying.

Um, AI can make it easier for just about anybody to make content, to express themselves, and I think that’s a good thing, and AI is p- enabling professionals to pioneer new types of art and live performances, games, storytelling. People are often at the center of this new work, and I want to give you one last example. So, a composer named Ann Heggie used my software to make music that explores how we could design new rituals around death. She built this really beautiful collaborative musical instrument that requires eight people to move together to play it. She used machine learning to make the building of this instrument feasible, um, and it’s been performed in concerts, but also used at community workshops on grief and connection. And so I think this gives us an alternative vision for AI that’s quite different from the one being sold to us by the media and certain corners of the tech sector. Um, and this is what I want the future to l- to, to move towards. We need, do need to do work to mitigate the harms of AI, but this should be informed by our understanding of the positive things it can offer, and by how creators themselves want to use it.

John Donvan

Thanks very much, Rebecca. And that is a wrap on this debate, and I just want to say, as I listened to our two panelists have this conversation, I was just delighted to see how, uh, as robustly as you disagreed with one another, you addressed and one another civilly. You heard each other out. That’s the thing that we aim for at Open to Debate, that you are open, and that you are open to debate. So, Rebecca and Jonathan, thank you so much for taking part in this program with us.

Rebecca Fiebrink

It’s been a pleasure.

Jonathan Taplin

Thanks, Rebecca. It was fun.

Rebecca Fiebrink

Thank you, John. It really was. I really appreciated hearing your perspectives?

John Donvan

And I also want to thank our, our guest panelists, the journalists and others who asked questions, Chloe Stead, Jason Scott, Eileen Kinsella, for contributing some very, very interesting questions. I also want to thank our audience, all of you out there, for tuning into this episode of Open to Debate. As a nonprofit, our work to combat extreme polarization through civil and respectful debate is generously funded by listeners like you, and by The Rosenkranz Foundation, and by supporters of Open to Debate. Open to Debate is also made possible by a generous grant from the Laura and Gary Lauder Venture Philanthropy Fund. Robert Rosenkranz is our chairman. Clea Connor is CEO. Lia Matthow is our chief content officer. Alexis Pangrazi and Marlette Sandoval are our editorial producers, and Gabriella Mayer is our editorial and research manager. Andrew Lipson is head of production. M- Max Fulton is our production coordinator. Damon Whittemore, our engineer. Gabrielle Iannucelli, our social media and digital platforms coordinator. Rachel Baker is events and operations manager. Rachel Kemp is our chief of staff. Our theme music is by Alex Clement, a human being, and I’m your host, John Donvan. We’ll see you next time.