Automation is rapidly becoming commonplace in a range of ‘human’ job spheres. The scope of what artificial intelligence and machine learning can accomplish more efficiently than humans increases every day. Jobs that were believed to remain exclusive to humans — that involve creativity and artistic judgment — are already being invaded (or augmented, as will be highlighted through the course of this article) by machine learning and AI.
One of the foremost instances of machines creating human art is that of Rutgers University’s Art and Artificial Intelligence Lab (AAIL) (Chun, 2017). Here, ‘human’ is used as an adjective to refer to the art’s ability to blur the line between AI-made and hand-made. Through a special Turing test conducted by AAIL’s Professor Ahmed Elgammal, it was proved that the images produced by the program were indistinguishable from museum-grade canvases when judged by a human audience. At this stage, in the involvement of AI with the creation of art, the following question invariably arises — is art actually art if it wasn’t created by a human?
This question is the first step towards establishing whether a piece of art created by a machine that possesses ‘artificial’ intelligence can effectively be termed as an ‘original’ piece of art. In order to make this demarcation; a universally-accepted definition of art is much needed. However, there is no such single definition.
Art can be any entity or form that invites human reflection and contemplation, with no mention of the means through which the artwork was produced. Other definitions may underscore the importance of the means or of the human touch — an inherent, individual skill and creativity without the application of which a piece of art cannot qualify as art. The definition of art provided by the Dictionary.com, that is the expression or application of human creative skill and imagination, typically in a visual form such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power, effectively establishes the importance of the ‘human’ quality in art. There are numerous terms in this definition that make it innately human. For instance, the emphasis on ‘human creative skill and imagination’, imagination being the keyword, underscores the necessity of an individual’s contribution to the creation of art. Furthermore, it may be imperative for there to be a purpose or strong emotions or other similar motivations for the viewer to feel the emotional power from the artwork, which may consequently require a human source. However, the researchers at AAIL, through their aforementioned Turing test, have also proved that an input of human emotion is not required for the viewers to feel that a piece of art is evocative.
The researchers at Rutgers University have generated images from their ‘Creative Adversarial Network’ that not only emulates unmistakably human art; the program is also designed to allow for deviations from historically established styles while remaining in the limits of ‘art’. Through the results of the aforementioned Turing test, it was also discovered that the viewers often rated the computer-made pictures as more ‘novel’ and ‘aesthetically pleasing’ than the museum-grade pieces (Chun, 2017). The fact that a computer program can create paintings indistinguishable from those created by a human mind did indeed present implications for the future of the creative, but doesn’t the evocative nature of these very paintings set those dire implications in concrete?
Dave King, founder of Move 37, a creative AI company, does not believe so. Taking into consideration the current level of application that creative AI operates, he believes that it can only bring a limited perspective to artistic practice (Bonini, 2017). “They can only draw on what they’ve been trained on, whereas the human condition is expansive and broad and brings a lot more depth of perspective to it”, he said in an interview with ABC News, while speaking about the reams of data used to create artificial intelligence. Along with King’s arguments on the limitations in current AI, groups like NIPS for Creativity or Machine Learning for Art (ML4A) illustrate how the use of neural networks in the creation of art does not mean the end of human creativity. Garry Kasparov, world champion chess player, believed that his battle with IBM’s Deep Blue led to better human chess players (Schaeffer, 2017). Along the same line of reasoning, Lee Sedol, expert Go player, reportedly expressed a similar opinion after fighting with a machine in the AlphaGo documentary.
Although AI can produce images that are on par or exceed the level of a human artist’s comprehension of creativity, there does not have to be any replacement of human ingenuity. On the contrary, AI can and should be viewed as a tool to augment human endeavors in the field of creativity, as it does in any other field. Similar to the idea of a paintbrush, AI facilitates artistic expression by being the means to an end. AI is one of the inflection points in the evolution of artistic expression through the centuries. It may even lead to unforeseen domains of elevated human creativity with the partnership of a machine in the near future.
Chun, R. (2017, September 21). Artsy. It's Getting Hard to Tell If a Painting Was Made by a Computer or a Human. Retrieved from https://www.artsy.net/article/artsy-editorial-hard-painting-made-computer-human
Bonini, T., & Donoughue, P. (2017, August 11). Artificial intelligence and creativity: If robots can make art, what's left for us? ABC News. Retrieved from http://www.abc.net.au/news/2017-08-11/artificial-intelligence-can-ai-be-creative/8793906
Schaeffer, J. (2017, July 10). Technology Review. Garry Kasparov thinks deeply about losing to a machine. Retrieved from https://www.technologyreview.com/s/608136/kasparov-thinks-deeply-about-his-battle-with-a-machine/
Image Credits: Kuaibao
Subscribe to our weekly newsletters.
Get all our posts, blogs and video content via e-mail.