After a judge in Colorado ruled that another award-winning artist cannot claim copyright due to MidJourney use, a great debate has flared up about the use of artificial intelligence (AI) for artistic endeavors.
In many countries, copyright ownership is limited to humans, so anything created by machines, or “mostly by machines,” must be confined to the realms of the public domain.
While, on the one hand, this is certainly not stopping artists from using generative AI to create new and unique pieces of art, it begs the question of whether these products can be protected by copyright laws or not. Mainly when the datasets used to feed these same AIs include countless hours of human creations taken from other human artists without their explicit consent.
But before considering the more philosophical and social considerations, let’s look at current legal developments.
The Matthew Allen Case in Colorado
Matthew Allen is a digital artist who won the Colorado State Fair’s first prize for the ‘New Digital Artist’ Award in 2022. His artwork, aptly named Théâtre D’opéra Spatial, immediately enflamed a heated debate among other competitors about whether it was fair and legitimate to use AI in a competition — even more so when it won.
Other artists claimed this digital art was midway between plagiarism and automatically generated content but in no way something that could be ruled as a winning piece — or even just a legitimate artistic creation.
The author, however, answered the criticisms by maintaining he never lied about how the artwork was created and that the creative process behind it is not so different from manipulating an accurate picture with Photoshop.
Other artists back up his position by confirming that human ingenuity is still central to generating such an image. In contrast, others dismiss it by claiming that anyone with some knowledge of prompting can do it.
In this regard, a skilled developer or programmer might quickly become a better painter than a painter in the near future.
Whatever the opinions may have been at the time, one year later, a judge from Colorado has now ruled that the artwork cannot be copyrighted.
Why? The artwork is not “human” enough; as we said before, machines cannot hold any right to hold copyrights.
This is not the first time US courts have made a similar decision. However, a definitive criterion to determine how much “human intervention” is required to define an opera, a book, a movie, or a song as human-made has never been set.
How Ethical Is It to Use AI-Generated Art?
A substantial part of the issue stems from the ethicality of the alleged property theft involving feeding machines with data from human artists.
As we already discussed in another article (or two, or three), several generative AIs such as Stability AI, DeviantArt, and Midjourney have been sued for intellectual property theft after billions of compressed images were fed to the AIs to teach them how to “create art.”
According to the artists who filed the lawsuit, their art has been used to train the AI without their explicit consent, to the point that what the machines generate now is a chaotic collage of individual human skills.
Whether these lawsuits will succeed is up for debate, especially since most of the art is in the public domain already or the artists behind them have long since passed on.
Philosophically, the border between what’s ethical and what is not is a fine line in the world of fine art. It could be easily argued that, besides obvious plagiarism, every artist is copying someone else to some extent. From ‘imitation is the sincerest form of flattery’ to ‘good artists copy, great artists steal’, there is little new under the sun.
We can debate how much of modern music lacks originality and any semblance of artistic novelty.
So, how is “made by humans” sufficient to define creation as true art?
Defining how much AI-generated content could be defined as “art” is not a niche problem as it may look at first glance.
Not taking a definite stance on the matter only means it will be a matter of time before the greediest companies will start flooding the market with incredibly cheap, average-quality AI-generated content, with no real restriction or reason against it.
And that’s a problem that falls into the more general “AI is stealing human jobs” debate in a pretty relevant way.
Being replaced by AI is worrying all kinds of artists, leading to the current and prolonged Hollywood actors and screenwriters strike.
Even if AI-made artwork is not protected by copyright, the risk of abusing the relative simplicity of generating it may cause many people to lose their jobs.
However, low quality is still low quality, and the public doesn’t react positively to it even when humans create it. And when the bottom line is affected, it’s hard to believe that companies will still want to go down that route.
We have argued that the market will eventually find its balance, making it no less critical to actively regulate this environment.
By definition, art is an ephemeral, subjective force that can hardly be described. The line between mere craftsmanship and true art is very blurred; even our simplest tools, like a sword or a flask, can become a masterpiece when built by the right hands.
What ultimately determines what is art and what is not is other humans’ perception and reception.
And if we’re entering a new era where machine-generated paintings are as enticing, if not even more, as human-made ones, well, there’s only so much we can do.