As the jungles become more concrete and our conversations become more digital, intelligence has become more artificial and machines are being taught to learn. We encounter artificial intelligence (AI) every day: when we check our email, cash a check online, use social media, look up travel directions, or ask our preferred digital assistant to set our morning alarm (techemergence). In recent years, fear has been conjured up around the future capabilities of AI, and the potential for AI to render many occupations redundant. To alleviate this fear, many have often made the illusory point that humans still hold their unique potential for artistic creativity over AI. However, hobbyists and University Labs have been rendering this point moot.

​

In 2017, the Art and Artificial Intelligence Lab at Rutgers University published their recent findings where AI-generated art gave off the impression of being human-made more often than human-made are from the Art Basel (artnet). This was only two years after the lab developed an algorithm that could analyze any work of art and determine the artist, genre, and style. They later used Generative Adversarial Networks (GAN) to generate art artificially. In this system, one network makes an image, and the other scrutinizes it. With each attempt, the GAN becomes more adept at producing art in the style of a genre or style. Rutgers has produced a modified version, called a Creative Adversarial Network (CAN) which prioritizes variance, and produces art that doesn’t adhere to style, much like contemporary art (artnet). Rutgers then asked humans if they thought a given work of art (which came from the CAN, GAN, or Art Basel) was made by a person. The results were: CAN, 53%; Art Basel, 41%; GAN; 35% (artnet).

​

Sidebar: Robbie Barrat, a 19-year-old researcher at Stanford University used a GAN to produce nude portraits after supplying the GAN with 10 thousand examples. He also created a rapping AI after supplying it with lyrics from Kanye West’s discography (Artnome).

​

Any individual without a strong proclivity for computer science can play still with AI algorithms. Deepdreamgenerator.com features several tools, including one which can transfer the style of one image onto another. In other words, any personal photo could appear to be a Monet.

​

But visual art is not the only artistic where AI can mimic human creativity. Software can create simple background music. Ampermusic.com can generate music based on some input parameters, like mood, instrumentation, tempo, and duration. In 2017, Spotify hired someone who was working on developing music-creating AI at Sony (Scientific American). Sometime later, Music Business Worldwide discovered that many of Spotify’s ambient music playlists’ music were credited to composers and artists which did not exist (Scientific American). Spotify has claimed that the names are pseudonyms for people who were supplied with AI-based software tools (Scientific American). Some have calculated that if the music in question was made by real artists, royalties would cost Spotify somewhere in the realm of 3 million dollars.

​

As AI advances, several logical problems are becoming less and less of hinderances to progress. By the same token, highly advanced AI produces many more existential questions regarding the value of artificially rendered art, human creativity, and the role that scarcity and source play in big-a Art.