Artificial Intelligence, or A.I., is being used more frequently as time goes on. With so many pieces of software being “taught” how to be creative, it’s safe to say there’s a lot of overlap when it comes to A.I. and the actual humans who dedicate their lives to perfecting their crafts.
Before I continue, I want to make clear the distinction between some different types of artificial intelligence. The ultimate goal with A.I. is to reach the level of AGI (Artificial General Intelligence). An AGI system would be able to learn, adapt, problem-solve, and communicate completely on its own, without having to rely on predefined algorithms. In laymans terms, it’d basically be sentient.
AGI encompasses both “true A.I” and “predictive models” to some extent. Often when people are talking about predictive models, they refer to it as A.I, when it’s not.
True A.I., typically refers to systems that can understand, learn, and apply knowledge in a way that mimics human cognitive abilities. This type of A.I. often gets mixed up with AGI because of how similar and advanced they appear to be. True A.I systems need to incorporate predictive models to function, but predictive models themselves are not A.I.
Predictive models are not inherently bad. There are many predictive model systems that have some practical real world uses. ‘Linear Regression’ is a relatively simple one, that can be used to predict house prices based on features like size, number of bedrooms, etc. ‘Logistic Regression’ is another classification algorithm used to predict the probability of a binary outcome. It is commonly used in healthcare, for example, and predicts the likelihood of diseases based on patient characteristics.
AlphaFold, which can predict the structure of proteins based on their amino acid sequences, has had significant implications for drug discovery, and understanding diseases. There’s also IBM Watson, which analyses large volumes of data, answers questions, and provides insights across various industries, including healthcare, finance, and customer service. ChatGPT is also one you’re likely to be familiar with. It is a text based predictive model that allows you to prompt for code, poetry, or anything else text based, but it is just a predictive model.
Then there are the predictive models that “create” artwork, music, and poetry. Systems like; AIVA, which ‘composes’ classical music. There’s DALL-E which is capable of generating images or “artwork” from text prompts given to it by users. There’s even Adobe introducing its own “generative fill” feature to Adobe Photoshop. These are the ones that people have been making a lot of noise about in the past year, and the ones I have issues with.
From this point on in this blog, when I say A.I.s, I’m referring to those “creative” predictive models that people generally refer to, or think of as A.I.
I could go into detail about how damaging these A.I.s are, and how they tend to steal from real artists to smash together the pictures they claim were created with no outside help. I could dedicate this blog to all of my fellow creatives, because regardless of your medium, we all share the same distaste for this abhorrent stab at our passions. There’s a lot of conversation happening around ‘A.I.’ in general, and the work that ‘A.I. Artists’ produce, that’s for sure.
Over the years, there have been many ways that artists have their work dismissed. Usually by people who believe that they’re being lied to when we tell them how long a certain piece took, and how much skill and practice it takes to get to that point. For many traditional artists (that use copic markers, paint, or any other medium of that sort), for example, their work is often claimed to be traced, or fake in some way that only the accuser can spot. A recently new one, in the past decade or so, is claiming that their pieces have been printed out.
Digital artists, like myself, have had all sorts of claims thrown our way. From claiming that our work is easy because the computer does all the work for you, or that digital art isn’t real art, and that it doesn’t count because it’s not made with a pencil and paper. I’ve had people dismiss my work, solely because they think it didn’t take any real skill to do it, and that they could replicate it without trying.
The newest one to add to the pile now, is asking if my pieces were done by ‘A.I.’, and it’s grating a little bit.
I’ve had it a few times now, by many different people, both in the form of an accusation, and a seemingly innocent question. I can accept that in today’s day and age, to the untrained eye, it can be difficult to spot the difference. I am not upset at the question if it is genuine. I’d rather be asked than have the theory hanging in the air without me being able to defend my work.
What tends to sting, however, are the people who use that claim to dismiss my work completely, even if they know there’s a real person behind it. I spend many hours creating my artwork. All the pieces I have online currently have collectively taken years of my life to produce, and that’s not even taking into account all the pieces I’ve created over the years, that have got me to where I am with my art.
If we look at one art piece of mine in particular, like Discovering the Lost, for example; that piece took me months to complete. I’m not talking about the space in-between each time I sit down at my PC, I’m talking about actually sat down and working on it. If you squash all of that working time together, it took me an insane amount of hours to finish. Every single one of my main pieces follow that same pattern.
Then, you have the amount of time I’ve spent on creating a world within my work. Granted, I don’t have a lot of it written down or sketched out yet, but the amount of time, and thought that goes into how I can tie each piece together, and what species I want to create and feature, and what lore I want to reveal is definitely lengthy.
I have been drawing since I was old enough to hold a pencil. I have piles of books, and stacks of loose paper that are solely dedicated to sketches I’ve done over the many, many years I’ve been drawing. Eventually, my skills started developing digitally. I have several files on my computer filled with old art pieces, many of which I still tend to take inspiration from in my current work. I digress slightly, but the point is that it has taken me decades to get to where I am now with my work.
The people who use A.I. to create work and pass it off as their own, regardless of the kind of creativity they want to tap into, are trying to shortcut the entire process. Anyone can be creative, and I wouldn’t dare try to stop anybody from becoming a photographer, painter, musician, or dancer, but you cannot shortcut the effort and learning needed to get to a decent level at your chosen craft. Most of all, these predictive models steal from real artist’s work.
I love my art so intensely. When I spend too long without picking up my pen, I feel like a part of me is missing, and the only way to feel complete again is to lose myself in what I’m creating. My art is a fundamental part of who I am, and I love every second I spend creating it, whether that’s the lore side or the actual art side of it.
As an engineer and roboticist, true A.I. and AGI excites me. I think that it could have some really practical real-world uses, and that it’s something that should be explored (carefully – sci-fi makes this look like potentially dangerous territory).
As an artist, though, predictive ‘A.I.’ terrifies me.
Regardless of what kind of creative you are; a musician, a dancer, a painter, or any other kind of artist, being creative is more than being able to do it. It’s more than simply possessing the ability to pick up a guitar and play the right chords. Whatever song you hope to play with that guitar will never sound right unless you have the emotion and the passion behind it. Creativity has always walked hand-in-hand with emotion, and feeling.
Which is where predictive models falter.
A dancer will never look fluid, free and effortless, unless they dance with emotion. Art will never look or feel right unless there is passion and feeling behind the work. With anything creative, it takes time to learn. However, while all art is attainable for all humans, it must take dedication, time, and passion to make it worthwhile doing at all.
Being an artist of any kind is more than just the end result. It’s about adoring the process of actually creating. Encoding passion and emotion into art lets others draw it out at a later date.
I have a theory that because people are finding ways to shortcut the learning process of being creative, they don’t fully understand why many artists have such a huge issue with generative A.I. artwork. They see it as the same thing.
I’m going to finish this blog with one poignant example;
In 1989, a pop artist by the name of Keith Haring created one of his final pieces. This piece is only known and referred to as “Unfinished Painting”. Keith died in 1990 at the young age of 31. His passing was caused by AIDS-related complications, and that piece was a self-portrait of his unfinished life. It was left unfinished on purpose, to honour all the people (including himself), who died unfairly early because of the AIDS epidemic.
A few months ago, a Redditor decided to have predictive A.I. “finish” that piece. If we ignore the blatant disrespect and clear lack of understanding behind this piece, and art in general, the “finished” piece has no feeling behind it. Without having seen the real work, you can clearly see where the A.I. takes over. The patterns make no sense, and exactly like I said above; there’s no feeling, no emotion, and no artist.
The original was already completed. It doesn’t matter that there’s most of the canvas still showing, or that you’ll never see what the finished version looks like. You were never supposed to.
I’m going to say that again; You were never supposed to. The original piece is a visual exampl of Keith’s unfinished life. A life he knew he’d never see the result of. This artwork shows that emotion in a truly beautiful and heartbreaking way.
Plus, the ‘completed’ version was only done with predictive A.I.. Put the original through A.I. again and it’ll give you a completely different prediction. A.I. will never be able to predict how Keith would have finished it.
When AGI is finally achieved, and we have sentient machines living alongside us, that are capable of creating something only they could come up with (like human artists do today), then I will have no complaints about machines making artwork. Right now, though, these systems and the people who use them, do not properly comprehend what art actually is. It’s more than pattern recognition. It’s often a lifetime’s worth of dedictation and emotion.
So, to conclude; no, I do not use A.I. in any of my work.
Cerys, out!