There was recently an interesting article here, about Artificial Intelligence recognizing creativity.
Researchers at Rutgers University have developed an algorithm that can assess the creativity shown in a painting and accurately predict the level of value our society has placed on it.
The big statement this is making is that we can quantifiably measure creative works. The article above used the criteria of “originality” and “lasting influence.
As a side note, I should mention that while these criteria are fine for measuring the works of great artists, they do not apply to the everyday creativity we need to elevate our lives. An extremely creative solution or product will never have “lasting influence” if only one person ever sees it. If these two criteria only are used, it is missing something extremely important that is often part of the definition of creativity: appropriateness (Sternberg, 1999, Handbook of Creativity). No doubt the items put into this programme are selected by humans – thus, we are still using human thought to select the appropriateness of something.
Moving on, one of the natural conclusions from this algorithm, as exemplified by Google’s Deep Dream project, is that once a computer learns the criteria to understand something “Creative”, it could theoretically make something creative. Deep Dream can identify objects in an image, but then it goes further: when it sees something that looks similar to an object, it changes it to look more like that object. See an example (more here):
Before we get too science fiction, I want to highlight a realistic application of this. If we can catalogue our ideas, or our processes or our products, we can analyse this catalogue to identify the areas that we may consistently be overlooking. With the aid of a computer database, we can design a system that prompts us with places to look for different and new ideas.
Developments such as this are fascinating for the potential creation of works and also their contribution to the advancement of computer science and AI, but a seriously beneficial application could be the analysis of the creative processes of humans. That is to say, not judging the final product, but helping us get there better.
Feature image courtesy of William Hook