AI Art and Creative Commodity in Competitive Markets

Ada Lovelace, who in the mid-19th century made contributions to the theoretical work of Charles Babbage, understood early on the underlying thread connecting the humanities to scientific discovery. What society perceived as two mutually exclusive minds, Lovelace believed to have a great deal in common.

Whereas art holds a mirror up to the world from which we can glimpse ourselves, science provides the language from which the universe can be understood.

Charles Babbage’s theories centered on the creation of what he called an “Analytic Engine,” using proofs to demonstrate how machines could be made to complete tasks based on certain algorithms. Lovelace saw the vast potential that mathematics had as a language to express truth. In her writings, she described this phenomenon as “poetical science.”

The accelerated development of artificial intelligence shows the actualization of poetical science in the 21st century. Special algorithms use text commands to produce original works of art. The popularity of the Lensa app at the end of 2022 has solidified AI art’s place in the mainstream.

The future has entered the chat.

Algorithms Can Program Computers to Mimic How the Human Mind Learns

Computers have the ability to learn patterns from data which they then implement when processing future data. Even the simplest algorithms can mine patterns and results from data that then perform more complex operations. However, these processes don’t just happen on their own without prompting. They require specific parameters set by algorithms that engineers develop to give the computer instructions to categorize, to identify patterns, or to perform mathematical operations.

Artificial intelligence is an umbrella term that encompasses a series of related automated process that come about as a part of mining data. Computers have the ability to learn from past data, all without direct human input. Through automation, machines can independently operate and become more sophisticated.

The algorithm used to create AI art is called a generative adversarial network, or GANs. First developed in 2014 by researchers, GANs work by taking the generative process in opposition to discriminative learning models.

Generative learning models function by predicting the probability of one factor given a certain condition. These algorithms then go through hundreds of iterations that compares new raw data to “real” data it received in training sets. Each time the data goes through the generative network, it produces a result that more closely resembles the true data, as the algorithm runs comparisons against false data.

This type of classification produces content that most closely resembles the parameters input into the algorithm. The sets of “training data” come from a multitude of sources which include photos and images as well as artworks that already have been created at any point in time. The images train the algorithm to generate results with more accuracy.

This presents a dilemma for owners of intellectual property. Especially as artists who legally own their original work might claim that machine learning algorithms have to steal from them.

The Existential Implications That AI Art Imposes on Artists

The rising popularity of AI art raises many ethical concerns. It raises questions regarding how to define creative ownership and what copyright law considers to be intellectual property.

Recently, the U.S. Copyright Review Board decided that AI programs could not be legally granted intellectual property rights for work they produced. This ruling came about from an attempt to register a copyright in 2018 to an AI entity called Creative Machine, which is owned by Dr. Steven Thaler.

Dr. Thaler’s claim fundamentally rests on the very idea of “poetical science,” as through his programming specific commands to mimic different psychological states, the works resulted from real creative expression. Essentially, he built his case by arguing that computers using automated technology would produce original, one-of-a-kind works on their own. Based on this logic, it’s the coding itself that produced variations that could not be replicated.

Copyright refers to original works produced on a fixed, tangible medium. By fixed, the work itself stays relegated to a physical space, like on a canvas, on film, or stored on a computer file. Copyright laws exist to protect the rights to ownership for art someone made that exists in a physical or digital space. The finished project is considered by law to be the intellectual property of its creator.

According to the reporting of the issue in ARTnews, the Board of Review ruled against this argument based on the claim that copyright law inherently only refers to creative works made by humans. Incidentally, the current language of U.S. Copyright Law grants ownership to the “author(s)” of a creative work, but as of yet the legal code lacks specific language that designates that the author has to be a human mind.

For this reason, I don’t believe that this legal precedent will hold up as the demand for AI-produced art and its existence in public spaces becomes more common. Advanced engines and highly sophisticated algorithms will make it harder to challenge a non-human creator’s originality as AI gets better at producing complex works autonomously.

This could be potentially threatening to artists who have to compete in the market place with machines. But the bigger concern lies in the potential for computer engines claiming creative ownership rather than the human mind who created the code. Profits would then go to the company who owns the computer engine. The idea of a non-human entity having the same rights as a living person isn’t so far into the future as we might think. Because this legally already happens to let corporations act as individual entities, with the same “rights” as a person.

Artificial intelligence poses threats in its potential to be used for corrupt purposes by people or organizations who only want profit. To prevent the future exploitation of human-generated artworks, the sins of the current art market, especially for smaller less-established artists, must be addressed now before taking shape in future technological advancements.