AI Art is a Threat to the Economy, Not Art (Or: The "Solution" to AI Art is Art Education)

Recently there’s been a surge of interest in artworks produced by “Artificial Intelligence” - and the fears that it may eventually end up “replacing” working artists since people would no longer have to commission art from human artists anymore. To the point where some people on GoFundMe are raising legal funds “in defense” of artists out there. Here are their demands:

  • Updating laws to include careful and specific use cases for AI/ML technology in entertainment industries, i.e. ensuring no more than a small percentage of the creative workforce is AI/ML models or similar protections. Also update laws to ensure artists Intellectual Property is respected and protected with this new technologies.

  • Requiring AI companies to adhere to a strict code of ethics, as advocated by leading AI Ethics organizations.

  • Requiring AI companies to work alongside Creative Labor Unions, Industry coalitions, and Industry Groups to ensure fair and ethical use of their tools.

  • Governments hold Stability AI accountable for knowingly releasing irresponsible Open Source models with no protections to the public.

While not without good intensions, these “laws” are unenforceable and will probably not get the results they’re looking for, even if passed. The reason is because in its current form, artworks produced by AI platforms can only be derivative - all it’s doing is pulling and mixing data from existing sources, stitching them together in a quasi-surrealist/collage style (which is specific to the algorithm, btw), then spitting out whatever results it gets in a quasi-randomized way. The better way to think of AI Art is that it acts as a “mixer” of sorts - an aggregate or amalgamation of what already exists in its database already.

On its own, in other words, AI is unable to produce anything original because all of its source material can eventually be traced back to a “real” work produced by a human artist at some point. The real problem with AI Art isn’t the results it produces, but the fact that it does not give proper attribution to its source material. Mechanically we know that in order for AI Art to work, it does need inputs from human artists (which includes the developers designing the algorithms) at every turn. What’s missing is that link documenting where its source materials came from - which is an issue that tech companies have historically struggled with up until now.

AI companies are selling the dream, rather than the reality, in other words. There are a few simple reasons why AI Art isn’t as big of a deal as most people think:

  • Derivative and variant works are already a thing in existing art practices. Think of all the “me too” and copycat works that already get created on a regular basis. (It’s a complicated issue, but the issue itself is not new.)

  • Each algorithm is only capable of producing one type of “style”. The creativity comes from the developers working on it, which is as labor intensive as creating a work from scratch.

  • Related to the above point, when something gets popular people get bored and move on. The more AI Art becomes popular, what becomes novel is what is not that. The concept of novelty is always an ever-moving target that can’t be deduced to a formula. (There have been many attempts historically, but all of them fail eventually for this same reason.)

  • If you’re looking for “free” art, royalty-free and public works are already a thing. (And many of them fairly high in quality.) Trying to make art even cheaper is beating a dead-horse, bordering on exploitative.

  • Art is often used by people to try to bring status and prestige to themselves or their institutions - the easier it is for people to produce these artworks, the less special it becomes. (See above: moving target for what constitutes novelty.)

At the end of the day, the goal of AI art is to be “passable” - not to produce great, or even good works. Its marketing practices are primarily directed at 1) people who are too lazy/cheap to hire/develop their own skill/talent and 2) people who don’t know or care to have a “message” in their art. Again, the real danger with AI isn’t its output - it’s that people might take its results too seriously and embrace the race-to-the-bottom model we already have now. At best, it’ll produce a below-average result of patchworked art with nothing to say - at worst, it will be used as a tumbler for laundering copyright from the hard-working artists out there. The way we talk about these issues need to get better, either way.

A Better Way to Use Artificial Intelligence

The interesting thing about AI is that it comes at a time where cryptocurrency and blockchain technologies have largely stolen the spotlight of what people consider “next gen” tech - there is a way to combine these ideas into a process that may be both interesting and beneficial.

All that would require is for existing AI algorithms to use blockchain data as its source material, with greater transparency as to where its getting its data from and how it is processing it. It would be informative, educational, useful and most importantly - fair to the creators who have put in the work contributing to the results of the algorithm. Since on-chain data and materials are largely considered public entities, it shouldn’t be too difficult to make this adaptation since the technology to do so is already there.

Beyond the marketing hype you hear about AI all day - you don’t hear too many stories of people letting AI make the “big decisions” for them. The AI is often too wrong and unreliable right now to be used in anything serious and will stay that way for a very long time. Anyone attempting to use AI art for branding or to satisfy their aesthetic needs will get labeled as “kitschy”, “cheap”, “missed the point” - deservedly, so.

Will these things get better in the future with better technology? Only if we scope it in ways that clearly understands AI’s powers and limitations. (We should probably call it what it actually is - a real-time machine-learning model, tbh.) That whole idea of the AI being able to do “anything” was always untrue - we need to acknowledge the human sides of the technology more in order for the technology to reach its true potential. It is the only way it can be used for real-life matters, rather than the toy that it is right now.

There is also another simple solution: which is to get the public more educated on how to do proper aesthetic appreciation, which will let people make quick judgements about whether or not the art that they’re looking at is authentic or novel. It’s a rare skill that may find its niche in the emerging NFT market to come.

Subscribe to ryangtanaka
Receive the latest updates directly to your inbox.
Mint this entry as an NFT to add it to your collection.
This entry has been permanently stored onchain and signed by its creator.