By HANNA HIRAKAWA
You may have heard about specialized computers beating human chess professionals, but what about resurrecting Monet or Van Gogh? Today, the level of artistic abilities demonstrated by Artificial Intelligence (AI) programs are indeed mind-blowing. One such program called The Next Rembrandt produces fresh artworks with a Rembrandt-esque style, almost mistakable as an original. The program is fed digital visual information and trained to recognize the artist’s brushstrokes, palette, and other distinct patterns that characterize Rembrandt’s painting style, according to the official page. Essentially, the goal of the mechanical brain is to generate new pieces when it is given minimal instructions of a theme.
However, AIs aren’t only resurrecting great painters. Other programs like Emmy (EMI), which composes music, and Kimagure AI Project Sakkadesunoyo, which writes short stories, are also AIs being put to test in the arts field. Additionally, Google Brain Team has developed Magenta, a project which aims to create an algorithm that produces “compelling art and music.” Moreover, the project is intended to connect experts from a wide variety of fields including coders, researchers, and artists during the process, according to Douglas Eck, a research scientist working on Magenta. Taking all of this in mind, the potential skill of machines in the arts seem to be progressing; and the conventionally perceived gap in artistic competency between humans and machines closing. Or is it?
Rembrandt brought back to life. The Next Rembrandt software processing programmed patterns onto canvas.
Photo by ING. http://creativity-online.com/work/ing-the-next-rembrandt/46306.
Much debate circulates when we begin pondering the idea of technology authoring artistic works, independently of humans. Are the systematically produced works by these machines truly art? In answering this question, one crucial criterion is originality. Originality emcompasses creativity and authenticity. When one appreciates Van Gogh’s paintings, we acknowledge the distinctness of contrasting hues and heavily rough brushstrokes. This set of visual features is what separates the artist from his contemporaries of the Impressionist Era and distinctly signifies his. And furthermore, these differences do not simply derive from studying differences and simultaneously gaining inspiration from existent artworks. Distinct styles develop from each artist’s unique perception of the real world, emotional experiences, and even personality. Currently, AIs’ creations is heavily based on the systematic analysis of existing past works, as is evident from The Next Rembrandt. In other words, although the content itself is entirely new, there are no intended themes or messages behind the piece. Hence such works can be described as merely rearrangements of used color and textural structures, lacking in meaning or emotional drive. And so if we were to define art based on originality in these terms, computer-produced works wouldn’t make the qualifications for “true art.”
However, the meaning of art itself seems to be evolving, emerging out of the conventional boundaries. Professor Simon Colton, the founder of a creative painting software called The Painting Fool, believes in the potentials of machines in the arts field. He strongly believes that robot minds are capable of exhibiting creativity, equivalent to that of humans. Interestingly, his idea of art transcends the artist’s body. “You can imagine a child doing exactly the same thing as software,” Colton explains. “[B]ut people project creativity onto the child and not onto the software, because of the context.” According to him, just as the definition of creativity leads to discussion, the meaning of art is highly subject to debate. He sees this as an opportunity to broaden our scope; he approaches the idea of art challengingly as he continues working on the development of The Painting Fool, which can now self-criticize, learn from past failures, and independently set itself a goal, and display its own personality.
The Dancing Salesman Problem. The Painting Fool software applying motional patterns of the human body into fluorescent scenes. Image by The Painting Fool software.
https://www.newscientist.com/gallery/painting-fool/.
Colton’s thoughts may be right. As AI technology increases its level of autonomy, we will marvel at the human-like machines and, simultaneously, may feel a bit disturbed. Beginning with Ray Bradbury’s There Will Come Soft Rains, the non-human advancing to the human level in performing skills and eventually replacing humans has been a repeatedly addressed theme in culture. Today, the concern feels ever more close as movies depict these machines in human form. If machines can do the same, why need humans to do it? This is a jarring question, especially when it comes to the arts, which many may have long assumed to be human’s specialized field. The answer is not an easy. However, what we do know now for certain is that it’s still a long ways before machines develop emotions and cognition based on real life experiences. And we can recognize the value of human-made art based on the emotional energy and drive that compels us to express our inner selves.