What does the future of AI in art look like? In our latest industry report, we predicted that AI will continue to develop exponentially into 2020. As the year has progressed, this has turned out to be true. With many industries starting to turn to AI-based solutions to solve their problems, the art world has begun to follow suit. This trend, however, raises new questions:
Can we still say that there is a creative act on the part of AI? What creative spark or impulse can be said to exist in a program? Does AI have agency or can it act on its own volition?
As has been proven time and again, we never truly know what the future holds. Every aspect of our world has been changed by the advent of COVID-19. The ever-expanding sphere of AI art is no different. The advantage of the art form is that it can thrive under any circumstances. Furthermore, both AI art output and visibility are going to increase as a result of more people surfing the internet in the confines of their homes.
To give you a sense of what we mean when we say AI art, let’s imagine that you have an algorithm, and that some of the inputs in that algorithm were images created by other people. Your AI algorithm receives the material that you, the programmer, provide, and out pops AI art you never imagined could be made. Have you infringed on the original artist’s work?
When you program AI to learn from a variety of inputs, you never know what will be the result. In the case of a work of art, writing, or music, who is the author of what emerges from an AI program? Is it the AI or is it the programmer?
In the case of AI art, originality can indeed be subjective. It is hard to find the exact point at which the initial artist’s work is no longer identifiable, but merely a part of a greater, conglomerated whole.
On March 4th of this year, Starta hosted "Frontiers in AI - the End of Human Art?," an investment trends & pop-up show, featuring art created using AI, along with more traditionally produced art.
The panel included:
- Gabriele Di Cerbo, Princeton University, Department of Mathematics
- Juan Bravo, Multi-Media artist based in Brooklyn New York
- Matthew Field, an attorney at Masur Griffitts + LLP
In this wide-ranging discussion, the group broached many of the questions previously mentioned. They also touched on the nature of creativity, intellectual rights, and copyright infringement, to name but a few.
Humanity has already been bested by machines in games, such as chess and Go. We need to be ready to accept the fact that art can now be created by AI. Much of it will be considered bad by most conventional standards, but certainly not all.
At the moment, much of computer-created art is being curated by humans who decide what is worth keeping and what deserves a swift “death” (we don’t know if it’s merciful or not). It is hard to say when we will reach the point at which AI will produce consistently competent works of art.
Human curatorship might itself soon fall to the sideline as greater strides are made in the realm of machine learning. Neural networks have begun to become more sophisticated to the point that a new type of network has emerged, called the "generative adversarial network" (or GAN), which can actually sift out poor output based on proximity to the original input. This “adversarial” component makes sure that whatever comes out has some kind of semblance to an object, or a concept, that a human can perceive.
“GANs have been used by many artists and scientists to produce new images from large datasets,” says Gabriele Di Cerbo, Professor of Mathematics at Princeton University, “Probably the most famous is Portrait of Edmond de Belamy (2018) that sold at a whopping $432,500 at Christie’s New York. Since then we saw a big improvement in the use of artificial intelligence in the artistic world. GANs are technically very hard to train but with the introduction of technologies like GoogleCloud and Amazon AWS, it is now easier to have access to these algorithms. This opened up an entire new world as more and more artists started using them.”
With all this in mind, can we still say that there is a creative act on the part of the AI? What creative spark or impulse can be said to exist in a program? Maybe a deeper question is, does AI have agency? This is a question being examined by AI specialists, Machine Learning experts, and philosophers alike.
The art world has been no less susceptible to this change. In fact, it has even been drastically altered. With all of us at home, including many AI artists, greater room is given to explore a variety of digital mediums, and a wider audience is now available to observe this latest stage of artistic development. The inclusion of AI in a myriad of industries and fields is not only inevitable but should be embraced as a welcome advancement and a definitive upgrade.
In our latest industry report, we predicted that AI will continue to develop exponentially into 2020. This has proven to be true, as we are currently witnessing a large variety of industries turning to AI solutions to complement, or even replace, their older ones. Industries that had once been perceived as stagnant have now been reinvigorated by machine learning technology.
As evidenced in our industry report, despite the seemingly dismal image outside of our own windows, the future of AI has never looked brighter.
To learn more about trends in AI read our report here.