Opinion mathematics from Oxford: can AI create as a person?
The game of go, in which the computer program DeepMind beat the champion among people, created a sort of confusion for Marcus du Sautoi, a mathematician and professor at Oxford University. “I always compared math with playing go,” he says. And go should not be a game in which a computer is so easy to play, because it requires intuition and creativity. Therefore, when du Sauthey saw how AlphaGo from DeepMind defeated Lee Sedol, he thought that there were changes in the field of artificial intelligence, which would affect other creative spheres.
The scientist decided to investigate the role that AI can play in our attempt to understand creativity, and wrote the book “Harvard University” in the Age of AI.
Artificial intelligence and creativity: who to whom?
The Verge and Du Sautoi discussed various types of creativity, how AI helps people become more creative (instead of replacing them), as well as creative areas in which AI faces the greatest difficulties.
Let's first analyze what “creativity” is, or artistic creativity. In the book you talk about three types of creativity. What is it and what does it mean for the role of AI?
Many people think that artistic creativity is an expression of what it means to be human, and if so, how can AI come closer to this? I look at many artists and show that quite a lot of works of art have a pattern and structure, which are very mathematical in nature. That is why I believe that artistic creativity can be more about patterns and algorithms than we suppose, and very often these patterns are hidden. Perhaps the AI can detect this, because it is very good at finding hidden patterns.
There is exploratory creativity that takes the rules of the game and takes them to the extreme, as Bach did. There is a combinatorial creativity when you take two ideas that have nothing to do with each other to see how associations in one can help stimulate new ideas in the other. The third creativity, which for some reason is the most mysterious, are those moments that appear as if from nowhere – something like a change of phases, when you boil water, water turns into steam and the state of matter changes completely.
How does AI fit into these schemes?
Each of these creative approaches offers different difficulties for AI. Research creativity seems ideal for a computer, because it is capable of producing much more computation than a human brain. Combinatorial creativity is interesting – AI can study patterns and apply them in new areas. But the most difficult for him to create something new and break out of the system.
Usually thought so: “How can AI break the rules? Is it not stuck in the system because it is programmed to work in a certain way? How can he jump out? ” But if the AI says: you must break the rules, this will also be the rule. You have a meta-code that tells the program to break the underlying code.
In your book, you talk a lot about creative AI projects. Which ones were especially interesting for you?
One of the most interesting was jazz Continuator, who took the music of a jazz musician, studied patterns and began to play on his own. The jazz musician’s response was striking. He said: “I understand everything I hear. This is my world of music. He plays just like me, with the exception of those things that I had never thought of before in my musical world. ”
Therefore, I think this is one of the exciting roles of AI in the future. People often begin to repeat patterns of behavior. Strangely enough, we are becoming more like machines, because we are simply repeating something, so it impresses me that jazz Continuator made the musician think a bit about his computer behavior. He helped to awaken his creative potential, showing that you can rearrange the ingredients that he already had, and he did not even think about it. I wanted to show that the role of AI in creativity, perhaps, is to increase the creative potential of a person, that this is a partnership of the future, that together we can make things more interesting than if we worked separately.
Another interesting story that, in my opinion, is important, is connected with the world of visual art and DeepDream from Google. Google has given a task to its visual recognition software to examine a random array of pixels and describe what they see. Through this, we learned something about how artificial intelligence was programmed and how it saw it.
What is the meaning of this?
One of the problems of modern AI is that many machine learning programs create code, but we don’t quite understand how it works. The Google DeepDream project helps us find a way to understand how this happens. Therefore, as for us – people – art is a way to penetrate into the consciousness of another person, perhaps the art created by AI will help to get into the essence of the work of this code, very mysterious.
Take the Microsoft Rembrandt project, which creates AI-generated images in the Rembrandt style. One could say: “Why do we need another Rembrandt? Have we not yet had fantastic rembrandts? ” The bottom line is that all this helps to understand the new in works of art. If you look at the work of Jackson Pollock from a mathematical point of view, we will see new things that we missed before. So AI can play an interesting role in discovering new structures that we may have missed in works of art and now take it for granted.
Such a search for patterns is not limited to the visual arts, right?
Well, in the world of cinema, you can take the algorithm Netflix, which recommends movies that we might like. He can share movies in interesting new ways. We could identify some of the groups as “all comedies together,” but sometimes the films are grouped according to the expression “like” and “dislike” by people, and then the general theme eludes us. It seems that AI has defined a new film genre for which we don’t even have a name. We can say that "there is a new flavor that you need to name." Perhaps the AI takes our creative works and sees in them something that we can express, but to realize – no. He could help us consciously formulate the essence of creativity.
There are many creative areas. What is the most difficult AI to do?
One of the surprises for me was how difficult it is to write words. Artificial intelligence has so much to write. I was quite surprised that even though the AI is writing quite well in brief, he is still unable to write for a long time. He does not have a good sense of narrative line, for example. I have not seen anything that would extend the coherent story beyond three pages. Perhaps the AI is very difficult to formulate language constructs as sophisticated as we are. Maybe he needs to go through the evolution that we have gone through. And then the question is: how long will it take?
You can answer this question in our chat in Telegram.