Well, I thought that I’d write yet another article about art and A.I. today. In particular, how it could utterly destroy one of the more interesting types of art learning, and what artists would lose if this happened.
This was something I ended up thinking about in early January when, after way too long away from the genre, I made a vaguely detailed semi-digital cyberpunk painting (using waterproof ink, watercolour pencils, a scanner and image editing software). Here’s a preview of part of it:
The full semi-digital painting should be posted here in early May next year.
Ok, in the weeks before this, cyberpunk art had of been relegated to something I made when I was feeling uninspired – a lazy “I can paint it in my sleep” type thing where I’d just paint a quick cityscape, the sort of thing that would just be a background for a more detailed painting. It’s a genre which is known and familiar to me, boring and ordinary. But there was a time – especially between about 2015-18 – when I was obsessed with making cyberpunk art.
Whilst I was a vague “on-off” fan of the cyberpunk genre beforehand, I think it really began in 2015. I’d either listened to Perturbator for the first time or I’d re-watched “Blade Runner” (1982) for the fifth or sixth time, but I was absolutely fascinated by cool-looking cyberpunk cities.
I wanted to explore the “world” of films like “Blade Runner”. I was so curious! But I didn’t have a huge amount of research material and I also wanted to make original art. And this was a major source of motivation for me. Whilst I explored it via art, I also frantically searched for anything even vaguely cyberpunk – which I could afford and/or which ran on the ancient computer I was using at the time.
Over those three years, I looked at a variety of different things – various second-hand “Ghost In The Shell” DVDs, a movie from the 1980s called “Trancers”, a DVD of the pilot episode of the “Total Recall” TV series, the “Robocop” films, the 2014 computer game “Shadowrun: Dragonfall“, the original “Deus Ex” (2000) etc… literally anything cyberpunk that I could get my hands on. I was curious and fascinated and I wanted to learn more and more, to learn the “rules” of what makes something look cyberpunk.
I quickly realised that I preferred “retro” cyberpunk – with its “film noir” rain and neon and bulky technology and Brutalist architecture and retro/vintage-style fashions – to “modern” cyberpunk.
Experiments with limited palette painting at around the same time eventually resulted in a more colourful style of cyberpunk art. I also refined my rule about how my art looks best when about 30%-50% of the painting is covered with black paint (to make the colours bolder by contrast).
My enthusiasm about the genre often led to these paintings having a slightly more joyous mood than cyberpunk media should probably have. As soon as I learnt how to add rain digitally, it showed up in almost every painting. As soon as I figured out how to alter the opacity of digital airbrushes, there was dramatic hazy “bloom” lighting etc…
I could go on but this eventually led to me developing my own “style” of cyberpunk art, mostly from research, from “working it out on my own”, from experimentation (and failure) and from curiosity. But, most of all, from exploration.
Back in 2015-18, when A.I. was a thing which only showed up in cyberpunk media, I had relatively limited source material, so I had to explore it through art. I had to imagine and wonder and research and guess and fail. All of this was driven by intense curiosity and fascination about a genre which I knew a bit about but wasn’t super-knowledgeable about. I wanted to learn it’s “rules” and, in the process, I came up with my own set of them. My own personal way of making cyberpunk art.
Of course – had I suddenly been curious about the genre in 2024 instead – then, with a few short sentences and in less than three minutes, I could have ended up with THIS:
Some 1980s-style cyberpunk images generated using DreamUp in January 2024. And, yes, I’ve added “A.I.” watermarks/labelling because A.I. images should be clearly labelled.
In a matter of minutes or hours, my curiosity would probably have been satisfied. And I’d have lost out on a significant part of my artistic development. Why spend years of practice and research learning the visual “rules” of a genre and, in the process, finding your own unique personal interpretation of it, when you can just follow the “efficient” path of least resistance and get a computer program to do it for you?
The reason why the old-fashioned way is better is “unique personal interpretation”. Yes, theoretically, if you created lots of variations of a single A.I. image, or carefully refined hundreds of prompts, you might end up with something a bit different… but that would just be one image. You wouldn’t be able to re-create it or really develop it further than that one image.
Having to study a genre of art, without easy shortcuts, leads you to learn more about your own sensibilities, your own interpretation of it. And I can’t help but feel that this sort of artistic development could be damaged by the proliferation of A.I. tools.
————–
Anyway, I hope that this was interesting 🙂
PekoeBlaze - the official blog 