AmImagine a world where highly trained and specialized robots display their high intelligence by creating unique works of art, all on their own. Perhaps in this world you will also find art exhibits with AI hybrid creatures interacting with museum visitors. Or immersive art journeys that take the viewer into a whole new world, guided by AI bots. Fancy? Futuristic? Science fiction? Not anymore. What used to sit squarely in the realm of science fiction has now become a pervasive part of our daily reality.
From smart home technology devices that sort your groceries to advanced search engine capabilities, AI is fast becoming an intrinsic part of contemporary life. The field of visual arts is no exception to that rule. While many artists continue to work in more traditional mediums such as painting, drawing, and sculpting with clay, other artists are embracing the new technologies available as technology continues to evolve.
In this article, we’ll look at specific examples of how today’s visual artists are using artificial intelligence to redefine art, including art created by AI bots themselves.
How AI contributes to art
Broadly speaking, one can categorize AI in any particular artistic process as fulfilling one of three roles: as imitator, collaborator, or artistic creator. Initially, in the early stages of machine learning development, artificial intelligences were trained to recognize specific image cues, learning a visual language of signs and symbols. Now, deep learning is more focused on language, to keep up with the demands of the Internet, such as keywords from search engines and social networks. But images are still an important part of the AI learning process.
As part of this initial and ongoing research, machines are trained to emulate works of art, acting as artistic mimics. This process is called “style transfer” and involves the machine studying and absorbing thousands of examples of a particular genre of art. By feeding thousands of images into a Generative Adversarial Network, or GAN, the machine learns to understand existing artworks and processes them through deep neural networks. The machine can then produce an image that would present a convincing appearance for such a work of art.
One artist, Robbie Barrat, used this process to create something new. By feeding thousands of nude portrait images of all styles, genres and time periods into a single GAN, Barrat helped guide the AI towards producing unique and stunning nude portraits. So for Barrat, the machine’s spoofing capability served as the ideal starting point for a human-machine collaboration.
Exploring the natural and the synthetic
There have always been artists who were quick to embrace the latest technological advances throughout history (the invention of photography and the development of portable paint tubes come to mind), and with AI this trend continues. Currently, a number of artists have chosen to work with AI in different capacities to explore the similarities and differences between natural and synthetic patterns, objects, structures, and modes of existence.
Anicka Yi is a great example. A multidisciplinary artist in the true sense of the word, Yi develops ambitious collaborative projects involving the expertise of scientists, chemists, engineers, AI specialists, and even master perfumers. Through these collaborative efforts, Yi seeks to open up and explore new modes of interaction and communication between AI systems and living organisms. In her recent exhibition, “In Love with the World,” Yi filled the Turbine Hall at London’s Tate Modern museum with floating robotic spheres inspired by oceanic and fungal life forms. These biotic forms were developed using sophisticated programming technology so that each was equipped with a network of possibilities for movement and interaction.
Jenna Sutela is another artist who combines biological, ecological, and artificial intelligences in her artistic practice. Her recent project has been a collaboration with Google engineers at Somerset House, also in London. Together, they are developing a “biocomputer.” This biocomputer is a machine that is being taught to channel non-verbal entities from the organic world, such as bacteria, and give them language. Through a collaborative machine learning process, Sutela is training these computers to extract language from a specific strain of bacteria called Bacilli Subtilis, a type of bacteria found in soil and in the human digestive tract. The combination of biological and artificial intelligences and technologies raises new questions about the possibilities of AI and its place in our world as technology continues to develop.
Surveillance, new realities and cautionary tales
While Anicka Yi, Jenna Sutela and Robbie Barrat are working with AI to explore new possibilities for creation and communication, other artists like Hito Steyerl and Ian Cheng are using AI to provoke audiences to question the status quo.
In Ian Cheng’s work, AI art is used as a starting point for viewers to reflect on the nature of art itself. How do we perceive art differently when we know it has been created by a machine? What will be the role of the artist in the future? And how can a viewer participate in a work of art that can be endless?
Using video game technology, Cheng creates works of art that are essentially video games that play themselves. With no way to actively participate, and with an endless series of small, inconsequential actions taking place over time, the viewer of these works always has the feeling that if they keep looking, they might find something important just around the corner. This process transforms the viewers themselves into passive observing machines, like the machines whose actions they are witnessing.
Meanwhile, Hito Steyerl uses AI to deconstruct and emphasize problematic political processes, such as pervasive state surveillance and artificial stupidity. Steyerl believes that AI profoundly impacts the way people think, and his videos and installations warn of the dangerous possibilities presented by widespread complacency. His works denounce the widespread spread of armies of bots that have the power to promote fake news and even influence election results. Working with AI allows Steyerl to immerse viewers in his political messages, with powerful results.
A future of embedded AI
With such widespread use and increasing familiarity in our everyday lives, intelligent machines seem to be here to stay. What the future holds precisely for the intersection of visual arts and artificial intelligence technology can only be told by time. There is a good chance that artificial intelligence will become another normal tool in the toolbox of many visual artists of the future. And it is possible that the AI bots themselves will continue to produce highly sought after works of art.
Regardless of the state of AI bot-generated artwork today and in the future, one thing seems clear. Given the innovative approaches artists are taking today with these emerging technologies, it seems likely that more collaborations between artists, AI engineers, and the machines themselves will continue to produce fascinating works of art for years to come.
This is a guest post by Josh Rogan, an expert in web technologies, Internet-age communications, and new AI developments. He believes that we have only just begun to harness the full potential of the Internet. Since he studied Computer Science, he has worked as an independent IT consultant, deepening his respect for what technology can do. He is now looking to try his hand at writing and share what he has learned with as many people as he can.
The main image in this post is AI generated by Disco Diffusion