Artificial Intelligence – A Creator of Art

“Analytical Engine (Computer) cannot create anything” – Lady Ada Lovelace

There is a group of people who believe in the said line of Lady Ada Lovelace, world’s first computer programmer. They firmly believe that computers lack the creativity needed to create something meaningful. A computer just does what it is programmed to do. Well, not going deeper into the discussion of computer programming being itself an art as argued by Turing Award Winner, Donald E. Knuth, we can clearly say that many of us do not agree with this school of thought.

Postponing the debate for a while, let us see what important role Artificial Intelligence plays in the process of creation of art. The role of AI can be broadly divided in two classes. First, they help artists in creating art works. This role is liminal in the sense that artists use the technology to create stuffs that are out of their capabilities and still are given credit for the same. The main argument behind this is that what is important is the vision or idea behind an art work and not the method. But can those ideas be realized without the available technologies. Take for example the software named Massive. It is used for crowd simulation in cinema for scenes ranging from wide-angle war scenes to a large stadium sequence employing Swarm Intelligence. Another example is of an animated movie from Disney named Tangled. It imitates the painting style of Jean-Honore Fragonard’s Rococo Paintings, specifically The Swing. Now we may say that the idea belonged to the creative director of the movie but could this have been pulled off without the use of a suite of highly advanced AI based software? The same holds for Massive. Could a scene comprising of millions of soldiers have been filmed without using the software? Here the AI tools create the art works and yet are devoid of the due credit. As Eric Guaglione, the creative supervisor of Disney says – “At Disney, it’s the computer programmer’s job to help an animated film’s director and art director realize their wildest dreams” and so true he is.

Second, they themselves create an art work. An argument that has been given for a long time in this context is that Computers lack feelings and emotions and therefore it is impossible for them to create an art work which is by definition a representative of feelings and thoughts of the creator. Here the question of consciousness while creating comes into the picture. Art work, as per my opinion, does not depend on the consciousness of the creator but on the consciousness of the audience. A painter may have been on drugs while painting a picture but this doesn’t mean he is not the creator of the art work. What would you call a child unknowingly humming a musical tune? Wouldn’t he be the creator of the tune even if he is not aware of the process of creation he is undertaking? We never know when we are creating a work of art. Same is the case with machines. As of now, machines lack feelings and emotions but they can create works that arouse feelings and emotions in the audience’s minds.

A great example of this would be a project by Harold Cohen and Kurzweil Cyber Art Technologies named AARON. It is a computerized turtle that paints pictures that have exquisite free hand ad-hoc look. It makes paintings that look like they have been made by a human and compels the audience to even think over the subject matter being depicted in them. The AARON turtle doesn’t even know what it has been painting. It just does what it has been trained (training is different from programming in the sense that we provide inputs to the system and allow the system to take decisions based on the training inputs whereas a programmed system does exactly what it has been programmed to do) for. I would definitely go for calling the intelligent turtle the creator of its paintings and not the turtle’s creators.

The brilliance of artificially intelligent systems is evident from another example which makes those who question the artistic abilities of AI systems look timid. It’s called Emily Howell. It is an artificially intelligent system that produces music on its own and a brilliant one at that. It even learns from the compositions of great musicians and creates new music in the same style. It has performed in several Operas and has released music albums of its own.

The creative power was realized way back by Alan Turing, one of the greatest computer scientists of the world which goes completely against what world’s first computer programmer had in her mind. Alan Turing was so confident of the creative power of computers and AI that he devised a method to test if an AI system is actually creative and at par with the creative prowess of humans, now known as the Turing Test. In simple words, a machine and a human is given a set of inputs and are asked to give answers to some questions based on those inputs. Another human is given the outputs of both and is asked to differentiate between them. If he fails to conclude which were by the machine and which were by a human, the system is said to have passed the Turing Test and is said to be intelligent. Some may be jumping the guns here, but Emily Howell is arguably said to have passed this test when people failed to conclude that the concert they just heard was by an Artificially Intelligent machine. However, there are people who have found the shortcomings in the test and as I have said, the debate goes on. For example, take an English-Chinese translator. It may very well translate the contents of one language to other and a normal human being may get confused, but it doesn’t understand a word of what it translated. There is no intelligence involved at all. The system may be said to be trained but it is definitely not educated. Just in the way, a speaking mynah or parrot doesn’t understand a word it is speaking but it speaks them because it has been trained to do so. But this argument goes against itself when we say that it is because of this reason that we require researchers with artistic sense and not just technical finesse to design intelligent systems and therefore, AI research, as it turns out to be, is itself an art.