The Battle Lines Over AI Art
Last month, artist Jason Allen won the Colorado State Fair’s art competition with a piece entitled Théâtre D’opéra Spatial. However, it was quickly discovered that the piece was not created by Allen directly, but was the product of an artificial intelligence (AI) platform Midjourney.
This prompted a sharp backlash against both Allen and the fair, with Twitter user @GenelJumalon putting it the most bluntly.
However, this wasn’t the only nor the most prominent contest that an AI won. At almost the same time, a short film entitled The Crow by filmmaker Glenn Marshall won the Jury Award at the Cannes Short Film Festival.
Though the controversy over The Crow was more muted due to the process Marshall went through to create the work, it represents another AI-created work that won a significant award that many people feel should be reserved for purely human creators.
But even as AI works rack up awards, they are finding themselves banned from an increasing number of spaces. According to a recent article by Benj Edwards at Ars Technica, several major art sites including Newgrounds and Inkblot have banned AI-created art from their services.
Others, such as DeviantArt, are quiet on the issue but are seeing an increased backlash from their users.
Both online and off, AI art is making its presence felt, and battle lines are being drawn around it. The question is simple: Will AI art find its place and, if so, what will that place look like?
A War on Many Fronts
AI art is controversial for a myriad of reasons. Many don’t see it as art at all, saying that art from machines lacks merit. Others, however, postulate that AI is just another tool, and that their creations are very much at the direction of human users.
But while the philosophical debates about what is art and whether AI art is actual art rages on, there are more practical and immediate concerns.
To that end, the battle lines over AI art are drawn over three separate fronts:
- Quantity: Since AIs can generate art much more quickly than humans, several sites have felt flooded and human creators can feel drowned out. For example, the furry art community Fur Affinity was dealing with one generator that was uploading an image every 40 seconds. This prompted a ban on AI works at the site.
- Copyright: All AIs are trained using existing images, almost all of which were created by human artists. Many artists have found their work in the libraries of different AI systems and have expressed anger over it. Though every AI is different in how it operates, some feel that AIs are not creating new works, but creating derivative works based on existing images. Whether that is true under the law has not been tested.
- Plagiarism: Though humans do direct AIs in the creation of art, humans are not doing the actual drawing, painting or creation of the work. As such, this creates questions of authorship and plagiarism when an AI work is published but not properly disclosed to be AI-generated.
In short, AI artwork has emerged so quickly that there are significant practical, legal and ethical issues surrounding it and the battle lines on all three are being drawn as we speak.
Unfortunately, all three are messy, and it could be years before norms are established. In the meantime, that sets the stage for chaos and a lot of hurt feelings.
Some Simple Steps for Now
Simply put, there’s no way to resolve all these issues right now. Even the legal ones are going to be messy and take years to get good answers on.
For example, even though the US Copyright Office has ruled that AI art can’t be protected by copyright, the divide between human created and computer created art isn’t a bright line, but a spectrum. As such, there are a myriad of questions about that space, even with a seemingly definitive ruling.
However, that doesn’t particularly help things right now and, as we’ve seen, many are already facing challenges with AI art already. That said, there are a few things that can be done today to help smooth things out.
- Always Disclose AI Art: If art is AI-generated, either in hole or in part, it should be disclosed. This is no different from disclosing the medium used or other works that were involved. Proactive disclosure prevents people from feeling misled by the artwork.
- Provide a Space for AI Art: Platforms that host art should make it easy for artists to tag their work as AI generated. A simple toggle to indicate if the work was created manually, partially by AI or entirely by AI would not only help prevent confusion, but also help categorize and organize the work. This could be bolstered by systems that detect AI artwork, to help spot cases where the usage wasn’t disclosed.
- Set Reasonable Limits: Platforms need to set reasonable limits on uploads. An AI that uploads an image every 40 seconds is no different from a human spammer. AI does provide new challenges for platforms, but TOS changes, rate limits and other changes can blunt the worst of those challenges without singling out AI art or artists.
While these can help address both the practical and ethical issues with AI, at least for now, it doesn’t do much for the legal.
However, there is a problem there. Though no court has ruled directly on the use of copyright-protected images to train an AI system, recent ruling such as the Authors Guild v. Google ruling make it seem unlikely that courts would rule against AI in this capacity.
In that case, authors objected to their books being scanned and used by Google to make a searchable database of books. However, the courts found that the use was transformative enough, and the amount republished was small enough to rule in favor of Google.
Ideally, the creators of AI systems would rely solely on opt-in sources or, at the very least, allow artists the ability to opt out. However, if a case does make it to court, it will be an uphill battle for the artist unless it can be shown that an AI-produced work copied protectable elements from the original work.
That is already a difficult hill to climb in many cases and will only get more difficult as AI gets better at hiding it “inspiration.”
Bottom Line
In the digital age, technology has a way of far outpacing our legal and social norms. We’ve seen this time and again as human systems try to keep up with digital evolutions.
AI is no different there but, depending on the path it takes, it could be a much bigger landscape change and come much faster than anything before it.
The best thing we can do right now is to try and be honest about it. It may be true that, someday, AI art will be just as accepted as oils paints or photography is today. As such, disclosing the use of AI may, at some point, seem quaint.
But if artists want communities to make room for them, they have to know who and what they are making room for. The discussion about whether AI art is art and where it should be allowed can’t be held if no one is clear on what is what.
Right now, transparency is possibly the most important thing, even if it feels like it isn’t necessary at all.
Want to Reuse or Republish this Content?
If you want to feature this article in your site, classroom or elsewhere, just let us know! We usually grant permission within 24 hours.