Traditionally, it had been believed,”How can AI break the rules? Isn’t it stuck inside a system because it’s programmed to do work in a way that is particular? How can it leap outside?” However, when an AI is told,”that I”m going to break the rules,” this is a rule in itself. You have a meta-code which tells the app to break the code.
This interview has been lightly edited for clarity.
Do not we have fantastic Rembrandts currently?” There is a point, and that’s that it could help us understand new things in the work of art. We can observe new items that we missed if you have a look in Jackson Pollock’s work from a mathematical standpoint. So there’s an interesting role in revealing new structures which we might have missed that we take for granted that AI can perform.
There are many creative fields. Is there one which you find AI and the maximum struggle?
There is exploratory imagination, which is taking the principles of the game and pushing them into the extreme, such as Bach. There is combinatorial imagination, where you’re taking two thoughts that don’t have anything to do with one another to see how relationships in one can help stimulate new thoughts from the other. The next creativity, which is somehow the most mysterious, are such moments that somehow seem to come from nowhere — those stage changes when suddenly you are boiling water, and water becomes steam and changes state completely.
In the area of film, let us choose the Netflix algorithm which urges the movies we might like. It can split films in ways that are fascinating. It identified a new genre of movie that we didn’t even have a phrase for. It is like saying,”there is another sort of taste inside here and you want to name this.” It see things that we’re expressing and can take our creative output but have not made aware yet. It can assist us articulate what might be there within our artwork.
The game of Go played between a DeepMind computer program and an individual winner made an existential crisis of sorts for Marcus du Sautoy, a mathematician and professor at Oxford University. “I have always compared doing mathematics to playing the game of Go,” he says, and Go isn’t assumed to be a game that a computer can play since it requires intuition and creativity.
Among the surprises for me was how challenging the written word would be. There is so much written sentence available for AI to find out. I was surprised that although AI is now quite good at writing short-form literature, it’s still not able to actually sustain the written word over a long term. It doesn’t have a great sense of story arc, for example. I have not seen anything that retains a story going beyond three pages. I’m really looking to see whether that is achievable, and I don’t see why it can not be, but it might be quite hard for AI to have the ability to articulate language as sophisticated as we do.
And where does AI fit into each of those routines?
So I think this is one of the roles that are exciting that AI will play going forward. Very often, humans start repeating patterns of behavior. It helped reawaken his imagination because he was revealed that there were things he could do using the components which he had and he hadn’t realized was possible. I wanted to show that the role an AI can perform in creativity is possibly to enhance human creativity, this will be a partnership going forward, that collectively we can make things more interesting than when we just worked independently.
We do not understand quite how it’s functioning, although one of the challenges of AI today is that lots of the machine learning applications produce code. The Google DeepDream project is helping us find a way to comprehend how that happens. So as art for us people is a way of helping us get inside the mind of the other individual, maybe art can help us get inside the workings of the code that is quite mysterious.
Let’s begin by breaking down”creativity.” In the book, you talk around three types of creativity. What are these and what do those breakdowns mean to the function of AI?
In the book, you talk about a great deal of creative AI projects. Which ones were particularly interesting for you?
Every one of these creativities offers another challenge to AI. Exploratory creativity seems ideal for a computer since that is what a computer can do, so many more calculations than the usual human mind could ever do. However, I feel the one is your concept of getting something and breaking from this system.
The Verge talked to du Sautoy about different kinds of creativity, AI helping people become more inventive (rather than replacing them), and the creative fields where artificial intelligence battles most.
Another interesting story that I believe is so important to this entire book is actually from the world of visual arts and that is Google’s DeepDream. Google requested its visual recognition software to see exactly what it saw in a random selection of pixels, and by dialing up the images it was seeing we heard something about the way the AI was thinking, how it had been seeing, and the way it had been programmed.
What’s the significance of that?
A lot of men and women feel that artistic creativity is about expressing what it means to be human, and so, how can AI get anywhere close to that? I look at a lot of artists and reveal that quite a great deal of art has pattern and structure behind it, which is quite mathematical in character. That’s why I think artistic creativity may be more about pattern and algorithm than we give it credit for, and very frequently the patterns are concealed, and perhaps that is something AI can detect because it is apparently somewhat good at discovering hidden patterns.
When du Sautoy saw DeepMind’s AlphaGo beat Lee Sedol, he thought that there had been a sea change in artificial intelligence that would affect other creative realms.
What was striking was that the jazz musician’s reaction. He said,”That which I hear, I understand. That’s my world of audio. It is playing like how I perform, except it’s playing things I have never thought of doing with my music audio universe.”