A very interesting interview at Spark with Michael Cook, a PhD student at Imperial College in the UK. It concerns the question of “artificial creativity,” which is basically artifical intelligence (A.I.) moving into the arts.
One of the most difficult issues surrounding this field, I think, is the way it raises the so what? bar even higher. It’s one thing to imagine the value of computers being “intelligent,” i.e. doing things for us so that we don’t have to do them and can concentrate on other things. But what good does it do for a computer to be “creative”? Is it to make them feel better about themselves? Or for us to feel better about them?
Certainly one of the values of artificial creativity is that it raises questions about the nature of human creativity. What good is creativity for in general? Are “creative” things (not just created things, let’s call them the arts for short) made for their recipients or their creators? Having computers create stuff for us to enjoy means we’re outsourcing the process of creativity at some level to our machines. It overprivileges reception over production. (You might actually think it’s cute for a computer to feel creative, in which case there is an ethics to artificial creativity.) In emphasizing the value of automating creation, we undervalue both the process of making something and the reception of that process that is partially based on our sense of, our communing with, another person’s creativity. When I read W.G. Sebald, to take just one example, I am not just enjoying what he has written. I am thinking about what it means for another person to have written this, that someone, both similar and different from me, has been able to think and write in this way. The arts bind us to each other in a very species-specific way.
But you could also say artificial creativity is valuable because it privileges a different type of creativity. Creating through A.I. is a way of creating creativity, but doing so in a different vein, one that can be based on features that humans do not have access to. In most cases, this means incorporating large amounts of information, or computation. We usually think of this as a means of devaluing the human (we’re not calculators! we love to say). But when an artist creates a painting, a composer a piece of music, or a writer a work of literature, they are each in their own way drawing on massive amounts of information, both experiential but also from the history of those individual arts. Our brains are amazing at integrating things, but there are of course limits to what we can take in, remember, and combine. Is there a value to being able to incorporate more into this act of synthesis that might be important to artistic thinking? Or is the point of art the excellence of the synthesis itself and not how much detail went into it? At the very least, we can see how talking about artificial creativity puts us squarely in the middle of debates about art itself.
And thenthere is the co-producer answer. The A.I. debate is a red-herring in this sense because it presupposes autonomy as opposed to cohabitation. We create computers that then create something, and in the process, we seem to forget that we created the computers (and the rules according to which they operate, or the rules that allow them to create their own rules, etc.). It seems to me that the real value of this kind of computational process — whether it is of the creative or intelligent kind — is that it aids us in doing things. There is a companionship to artificial intelligence/creativity. In this way, it is important for us to remember the artificiality of artificial creativity and not pretend like there is a naturalism out there where there isn’t one. Artificial creativity is just that: artificial. It’s a means, not an end.
Finally, as I write in my book, one of the most important values of these experiments in artificial thinking is the way they allow us to think about thought. In modeling, in failing, we learn more about how we succeed. I’ve always found this process to be incredibly beautiful.