Hey folks! Welcome to another discussion thread. This time, I'm wondering about the role of generative AI in enhancing or hindering creativity in the near future. Here's my question:
With all the amazing tools coming up for automating creativity (writing assistants, image generation, text-to-video, etc.), what is left for the human creator?
Will AI kill creativity? The short answer is no, if you mean the brain's capacity to be creative. This is a biological feature of language: creative use. We can utter novel sentences effortlessly, which is what Descartes found so perplexing. It continues to baffle many bright minds.
The creative faculty of the mind at a cognitive level "above" or "below" language is barely understood. As a demonstration, after thousands of years of philosophy, there is no agreement about what is art. It is arguably the most quintessential form of creative expression. Within the gamut of ideas of what is art, it is perfectly feasible to argue that what an AI produces is not art, or one can equally argue that it is and in some sense it has killed human-generated creativity.
There is a body of literature about Conceptual Blending, which also appears to be a biological feature of the brain -- synthesis. At this level of description, it is perfectly feasible to describe Generative AI as being creative. It can certainly generate novel outputs based upon blends. I recently asked it to generate a rap song in the style of Ni**as in Paris by Ye, but about AI. The output was novel in the first sense of creative use of language, but also novel in the second sense of conceptual blending.
Indeed, the AI appears to attend to the circumstances (my request) with something appropriate to those circumstances. Had I asked a rapper, lyricist or poet to do the job, the output would have been similar in concept: taking words from one domain (AI) and inserting them into the rhythmic meter of the rap -- i.e. the AI is being creative in human-seeming ways, yet far more productively.
To that end, if we are asking whether or not human creative outputs will decline relative to AI creative outputs, then this seems almost a certainty because AI tools are way more productive.
What is left for the human creator? If art is just aesthetic objects, then AI will soon generate far more than humans and nothing is left for those whose skill is producing similar objects to be used for similar purposes. So then we are back to understanding what is art and asking how will its many interpretations emerge and evolve in as post GenAI world.
Different responses will emerge depending upon whether or not the viewpoint is one of pushing back against AI (e.g. perhaps placing a higher value on rare objects, like pencil drawings, or authentic experiences like live music) or embracing AI as another tool in the creative palette.
In a sense then, there is nothing new here because if art is about mastery of skill (e.g. painting a landscape), then new artists will emerge to become masters of AI-assisted production. Consider that anyone can write a prompt and generate an aesthetically useful object, so writing prompts is not what makes a "Generative AI artist" an artist. It's just what makes a regular person an image producer.
So we are back then to the most human of traits: the imagination. Whoever conjures imaginative ways of using AI will become the new art masters. I am not sure it's the right metaphor to say "the bar has been raised", but in some sense it is a truism.
Love, especially the conclusion. Whoever conjures imaginative ways of using AI... In a sense reminds me of Knuth's definition for art. It is what we still don't understand well enough to explain it to a computer. It is a truism, as you say.
The bar will certainly be raised, at least in the short term. I can see this in my own articles. I’ve always wanted to write in my own style and not simply ape other Python tutorials, but now, I’m less likely to shy away from off-the-beaten-track ideas, looking at perspectives that are less obvious as generative AI can deal with those easily, anyway!
Yeah, I think that is a good point. AI can generate mundane content pretty well, which is a two-edged sword.
Having AI generate mundane stuff gives us more time to be creative. But also if the stuff we create is mundane it will get lost in a sea of AI content.
I guess it can be a baseline equalizer, lifting everyone to, say, the 70th percentile of quality. A lot of people will struggle to find a job, though. Especially those closely below or above the "good enough" line.
Creativity can't really be automated right now ... any creative requires human/machine interaction. A human cannot be creative without technology. Technology cannot be creative without humans.
Sure, you can automate generation ... and maybe a machine will randomly create something new, but current AI technology won't be able to recognize it. Kirby Ferguson, author of Everything is a Remix, made some good points about this recently.
"Computers can now create but they are not creative. To be creative you need to have some awareness, some understanding of what you’ve done. "
Now if you are a human "generator," it might be a different story.
"A human cannot be creative without technology" is an interesting one. Hadn't thought it like that before, but now that you say it, it really sounds like obviously true.
This is a "posthumanist" view. We are used to thinking about humans as the center of universe and the center of meaning-making. But meaning is created through interaction between technology and humans. So technology has always influenced how we create. Just think about how the printing press or internet changed the way we make meaning!
Yeah, exactly. Even cave paintings are examples of human-technology interaction. In a sense, even human language is just another massively successful technology.
Human creativity will likely move from the realm of thinking/reasoning within known problems or situations (this will become the domains of AIs) to imagination and emotional or irrational expression. Some of this creativity will fuel research and progress in science and some will be more conceptual and will define new ways to do art that are human specific.
Interesting take. Do you think art is by definition human? Seriously asking, as I've seen arguments for both positions. If you define art in functional terms (it provides such and such value) then of course eventually AI art will be at least as good as the best human art and maybe even better. But if art is intrinsically human, e.g., if you have to factor in the purpose of the artist, then maybe by definition generative AI will be something different altogether.
I do think art is intrinsically human. Anyone can now produce pictures that are functionally as intricate and complex as old school masterpieces.
But do they communicate anything to the perceiver? In the majority of cases, I don’t think they will. Because we, as art perceiver, will get used to this esthetic richness and will come to expect more from human directed art: some codes or hints in the work itself that communicate directly to us, human to human.
I view it as a "standing on the shoulders of giants" kind of situation: I'm hopeful that people will be empowered to realise their creative visions, though I do think there's a risk of people losing the value of creative *skills*, in the same way that people don't necessarily know how to build or maintain a car engine.
So I think ordinary people will become more able to take an idea through to execution, but may not learn, or even understand, what actually goes into it, and therefore may not be able to manually execute their vision themselves. That isn't necessarily a terrible thing however, I just think it's something which some will value whilst others don't.
There is a long tradition of tech allowing people to do more while understanding less. I'd even say that's tech's thing, and I'm not sure is a bad thing. What I worry about gen AI is people loosing interest in human art. It becoming so elite that only the connoisseurs or the rich can afford to read a poem actually written by a human.
I would make the case for two different "definitions" of art, so to speak. On one side, art, as a consumer, is somewhat elitist in the sense that if I can't satisfy my most basic needs I won't even consider buying or otherwise consuming artistic "products". But on the other hand, as a creator, I think one can view art as a means for expression (at least genuine art, but that's even hairier) whose value for the maker is independent of who ends up consuming it. I think about fanfic writers, street performers, small time YouTubers, i.e, people who make art just for the sake of expressing themselves, and the economic relationship is secondary if it exists at all. That's the kind of art I expect to be much harder to replace by AI.
Will AI kill creativity? The short answer is no, if you mean the brain's capacity to be creative. This is a biological feature of language: creative use. We can utter novel sentences effortlessly, which is what Descartes found so perplexing. It continues to baffle many bright minds.
The creative faculty of the mind at a cognitive level "above" or "below" language is barely understood. As a demonstration, after thousands of years of philosophy, there is no agreement about what is art. It is arguably the most quintessential form of creative expression. Within the gamut of ideas of what is art, it is perfectly feasible to argue that what an AI produces is not art, or one can equally argue that it is and in some sense it has killed human-generated creativity.
There is a body of literature about Conceptual Blending, which also appears to be a biological feature of the brain -- synthesis. At this level of description, it is perfectly feasible to describe Generative AI as being creative. It can certainly generate novel outputs based upon blends. I recently asked it to generate a rap song in the style of Ni**as in Paris by Ye, but about AI. The output was novel in the first sense of creative use of language, but also novel in the second sense of conceptual blending.
Indeed, the AI appears to attend to the circumstances (my request) with something appropriate to those circumstances. Had I asked a rapper, lyricist or poet to do the job, the output would have been similar in concept: taking words from one domain (AI) and inserting them into the rhythmic meter of the rap -- i.e. the AI is being creative in human-seeming ways, yet far more productively.
To that end, if we are asking whether or not human creative outputs will decline relative to AI creative outputs, then this seems almost a certainty because AI tools are way more productive.
What is left for the human creator? If art is just aesthetic objects, then AI will soon generate far more than humans and nothing is left for those whose skill is producing similar objects to be used for similar purposes. So then we are back to understanding what is art and asking how will its many interpretations emerge and evolve in as post GenAI world.
Different responses will emerge depending upon whether or not the viewpoint is one of pushing back against AI (e.g. perhaps placing a higher value on rare objects, like pencil drawings, or authentic experiences like live music) or embracing AI as another tool in the creative palette.
In a sense then, there is nothing new here because if art is about mastery of skill (e.g. painting a landscape), then new artists will emerge to become masters of AI-assisted production. Consider that anyone can write a prompt and generate an aesthetically useful object, so writing prompts is not what makes a "Generative AI artist" an artist. It's just what makes a regular person an image producer.
So we are back then to the most human of traits: the imagination. Whoever conjures imaginative ways of using AI will become the new art masters. I am not sure it's the right metaphor to say "the bar has been raised", but in some sense it is a truism.
Love, especially the conclusion. Whoever conjures imaginative ways of using AI... In a sense reminds me of Knuth's definition for art. It is what we still don't understand well enough to explain it to a computer. It is a truism, as you say.
The bar will certainly be raised, at least in the short term. I can see this in my own articles. I’ve always wanted to write in my own style and not simply ape other Python tutorials, but now, I’m less likely to shy away from off-the-beaten-track ideas, looking at perspectives that are less obvious as generative AI can deal with those easily, anyway!
Yeah, I think that is a good point. AI can generate mundane content pretty well, which is a two-edged sword.
Having AI generate mundane stuff gives us more time to be creative. But also if the stuff we create is mundane it will get lost in a sea of AI content.
I guess it can be a baseline equalizer, lifting everyone to, say, the 70th percentile of quality. A lot of people will struggle to find a job, though. Especially those closely below or above the "good enough" line.
Creativity can't really be automated right now ... any creative requires human/machine interaction. A human cannot be creative without technology. Technology cannot be creative without humans.
Sure, you can automate generation ... and maybe a machine will randomly create something new, but current AI technology won't be able to recognize it. Kirby Ferguson, author of Everything is a Remix, made some good points about this recently.
"Computers can now create but they are not creative. To be creative you need to have some awareness, some understanding of what you’ve done. "
Now if you are a human "generator," it might be a different story.
https://www.freethink.com/robots-ai/ai-and-future-of-creativity
I loved Everything is a Remix!
"A human cannot be creative without technology" is an interesting one. Hadn't thought it like that before, but now that you say it, it really sounds like obviously true.
This is a "posthumanist" view. We are used to thinking about humans as the center of universe and the center of meaning-making. But meaning is created through interaction between technology and humans. So technology has always influenced how we create. Just think about how the printing press or internet changed the way we make meaning!
Yeah, exactly. Even cave paintings are examples of human-technology interaction. In a sense, even human language is just another massively successful technology.
Human creativity will likely move from the realm of thinking/reasoning within known problems or situations (this will become the domains of AIs) to imagination and emotional or irrational expression. Some of this creativity will fuel research and progress in science and some will be more conceptual and will define new ways to do art that are human specific.
Interesting take. Do you think art is by definition human? Seriously asking, as I've seen arguments for both positions. If you define art in functional terms (it provides such and such value) then of course eventually AI art will be at least as good as the best human art and maybe even better. But if art is intrinsically human, e.g., if you have to factor in the purpose of the artist, then maybe by definition generative AI will be something different altogether.
I do think art is intrinsically human. Anyone can now produce pictures that are functionally as intricate and complex as old school masterpieces.
But do they communicate anything to the perceiver? In the majority of cases, I don’t think they will. Because we, as art perceiver, will get used to this esthetic richness and will come to expect more from human directed art: some codes or hints in the work itself that communicate directly to us, human to human.
I view it as a "standing on the shoulders of giants" kind of situation: I'm hopeful that people will be empowered to realise their creative visions, though I do think there's a risk of people losing the value of creative *skills*, in the same way that people don't necessarily know how to build or maintain a car engine.
So I think ordinary people will become more able to take an idea through to execution, but may not learn, or even understand, what actually goes into it, and therefore may not be able to manually execute their vision themselves. That isn't necessarily a terrible thing however, I just think it's something which some will value whilst others don't.
There is a long tradition of tech allowing people to do more while understanding less. I'd even say that's tech's thing, and I'm not sure is a bad thing. What I worry about gen AI is people loosing interest in human art. It becoming so elite that only the connoisseurs or the rich can afford to read a poem actually written by a human.
Interesting point. One could argue that human art has always been for the elite 😆.
I'd be curious to know what "creative skills" are analogous to maintaining a car engine.
I would make the case for two different "definitions" of art, so to speak. On one side, art, as a consumer, is somewhat elitist in the sense that if I can't satisfy my most basic needs I won't even consider buying or otherwise consuming artistic "products". But on the other hand, as a creator, I think one can view art as a means for expression (at least genuine art, but that's even hairier) whose value for the maker is independent of who ends up consuming it. I think about fanfic writers, street performers, small time YouTubers, i.e, people who make art just for the sake of expressing themselves, and the economic relationship is secondary if it exists at all. That's the kind of art I expect to be much harder to replace by AI.