To give my credentials as they pertain to tech, I worked in the field of high-level server and network support for over a decade and a half. Not that it should give any gravitas to the work, but my clients included many of the Fortune 500. My credentials as they pertain to art go back even further. I’ve been a storyteller for more than thirty years. I also play multiple instruments, some of them better than others. I’m lucky enough to be married to a visual artist. My friend group consists mostly of other artists and musicians. This is all to say that I have a pretty good idea about the intersections between tech and art.
AI has always been an interesting concept to me. The idea of a human-made consciousness or the more likely emergent consciousness forming in synthetic neural networks is very compelling. Will it become smart enough to take out all of humanity or will it become a benevolent god that helps humanity ascend to the next level? Will our future be Star Trek or will it be the Matrix? These were the types of questions I liked to noodle on. Now I’m stuck noodling on how stupid the current application of what ‘they’ are calling AI actually is.
I’m not some anti-AI luddite mind you. There are plenty of applications where its usage has been or will be revolutionary. Understanding the human genome to make cancers and certain disorders a thing of the past, frame interpolation to make overworked animators’ jobs easier, or AI-assisted coding to produce optimized routines that help reduce resource utilization are all practical applications that do not stoke my ire.
I don’t even mind some applications in the field of arts. It has recently been used to aid in the creation of a new/old Beatles song. Say what you will about it but they had the buy-in of every person they should have and AI was simply a tool. Most new spelling and grammar checkers use it which means I’ve used it as a tool for editing stories. There are also novel uses for it in art restoration. Here’s the thing, in every case I’ve outlined AI is used as a tool and not a means. This is to say it is helping artists, not attempting to supplant them.
If you are reading this you probably already know but AI Art generators rely on training. The method by which this training is accomplished uses very unscrupulous means. AI researchers generally just wholesale scrape sections of the internet for images and text content. They did not receive permission from most artists or authors to do so. Adobe has what they are calling an ‘ethical’ AI system but they hid everything an artist would be signing on for under layer upon layer of terms of services. They also do not allow artists to revoke consent once they learn the truth. At the end of the day, the systems are essentially absorbing the essences of hundreds of thousands of artists and regurgitating homogenized amalgams that lack any nuance or soul. It’s Cronenbergian at its core. There are a lot of AI apologists/bootlickers trying to justify the ethics of stealing someone’s art as a means to supplant them. Even worse, there are people who have bought into the ease by which images can be produced and claim to be AI “artists”. Let’s examine a few of these “artist’s” arguments.
In my internet travels, I have seen many AI “artists” complaining that real artists are not giving them a seat at the table. Their “art” isn’t seen as valid. Putting the ethical considerations aside, one of the main problems here is that they even have the gall to call themselves “artists”. Their claim is usually that there is a lot of technique in forming the text used to create the image output, that their true artistry is in the creation of the text that gets fed into the system. What if I were to tell you there were these things called art commissions where people put together a block of text describing in detail what they want an artist to produce? This block of text then gets interpreted by an artist who in turn sends you back an image that may or may not be revised and reworked by further blocks of text. We don’t call these blocks of text techniques, we call it basic human interaction. And you know what else? The commissioners never call themselves artists or otherwise claim to possess artistic talent, that’s why they found an artist. It’s almost like the artists are the ones who actually possess the desired skill in these transactions.
Another argument I’ve seen is, “I’m just playing around with the tech for fun,” as they post an image to their social media timeline. This may be the most forgivable of the arguments as most people are naïve to the ways in which AI “art” is harming artists. Yeah, they may be producing images that they otherwise would not have commissioned if the tech didn’t exist. The primary problem with this, aside from the previously mentioned wholesale thievery, is that it raises awareness of the tech and encourages other people to use it. This kind of, “everyone else is doing it,” organic growth is how public perceptions shift. These are the kind of shifts where artists get lost in the mix. If your argument is that this is slippery slope thinking then I will point you to the music industry in the late nineties and early aughts when wholesale thievery had become mainstream. It nearly destroyed the entire industry. Yes, the record companies were the pinnacle of greed and that market was in need of a major disruption but this was not the primary consideration of the pirates. For the most part, they just wanted something for nothing and didn’t consider the artists in the equation. In the AI debate, you don’t have the corporation to scapegoat, you are actually harming artists directly.
Yet another argument I’ve seen: “You just don’t understand the tech.” Isn’t it odd that this was the same argument that the crypto bros used when trying to defend wasting megawatt upon megawatt of energy to mine something that doesn’t really exist? It’s odd that they mostly went away then a mere year later all of these AI “artists” showed up on the scene. It’s also the same hollow logic that guns rights activists use to to claim superiority when they say things like, you don’t know what the AR in AR-15 means so I’m not going to take anything you say seriously. You shouldn’t have to argue with red herrings to feel superior. And trust me on this, I’m probably going to know more about AI and the tech behind it than some random AI “artist” trying to argue why a system built on art theft is not unethical and wrong.
Let us not forget the ad hominem “argument” of, “You’re just jealous that a computer can do art better than you.” This attack is just a logical fallacy and by nature already invalid.
And my favorite argument is, “How is it any different than an artist working with collage, a hip-hop producer sampling beats, or a photographer taking a picture of what is already there?” This is the real gotcha in the AI apologist’s arsenal, isn’t it? Not quite, you see in the case of both collage and sampling, the aim of the art is generally transformation and not homogenization. It’s my personal opinion that one is better than the other but I think you would find most artists would agree with me. They are taking what is there and turning it into something more than its base parts. You don’t really find that in AI “art” which is generally just a distillation. Some of the value in the art comes from the discussion surrounding whether it’s right or not to do certain transformations. You can look no further than the “Fearless Girl” statue which was presented in such a way as to be transformative of Wall Street’s “Charging Bull”. The creator of the bull had a problem with how the meaning of his work was distorted, a larger conversation ensued about the nature of patriarchal constructs in financial institutions and about whether the true meaning of the original piece was diluted by the new one. I have my opinions others have theirs. The “Fearless Girl” was eventually moved at the behest of the “Bull” artist, a move devoid of self-awareness. There have been legal disputes over the transformative nature of looping in hip hop and as a result, laws have been created whereby song credits, licensing, and royalties come into effect. Do you see how that is not the same as what is currently happening with AI art? If your excuse is that it’s okay because it’s still the wild west then you’re missing the point. As for photographs, there is way more that goes into a photo than simply capturing what is in front of you. Detractors can pretend not to know this but feigning ignorance is not ample reasoning for a bad take. Between proper lens selection, exposure timing, lighting, posing vs candid considerations, composition, and about a thousand other things, there is no way you can compare photography to AI in the way that most do.
There are many other arguments out there and I’m sure this post will drag a bunch more out of the woodwork. I’m actually kind of curious about the logical leaps I get to see next. Maybe I’ll do a part two at some point. To summarize the issue at hand, artists generally don’t want their works used for purposes they couldn’t have planned for when they created them. AI “artists” don’t want to feel bad about asking a computer to make them something shiny and devoid of meaning on systems whose creators either didn’t consider the ethics of what they were doing or worse, considered the ethics and wrote off any concerns, hoping they could sway public opinion against real artists later. AI “art” as it stands is unethical and should not be used. Maybe someday there will be a system by which artists are compensated properly for the usage of their art. Funnily enough, those images will still be soulless and lack artistic merit. But that day is not today, so can you please just stop?