23 June 2022

Phoebus

Phoebus

“Shimmering in reds, oranges and yellows, they watched as vast cloud banks, faced with angry storm fronts of blues and purples rolled across its surface, seemingly without ever dispersing. And the atmosphere, unusually crystal clear. There was something about it that captivated anyone who spent too long watching from afar.”


Experimenting with the AI driven Mid Journey engine.

When AI driven ‘art’ started to appear, I was interested from afar and then, as I saw more, started to think that this direction is just yet another thing that will put a tier of people out of work at some point in time. To a degree I still think that, and watching the speed it’s developing and how good it’s getting, anyone would be foolish not to think to. Sam Williams, a professional concept artist and modeller in the film industry had this to say about what he’s seeing. I can not disagree with his sentiment.

At the other end of this argument though is the position that these AI art generators are, and will be, another set of tools for the artist.

I decided to spend ‘quality time’ with Mid Journey (as opposed to slapping one’s way through generating some half baked image) using it to generate some images based on ideas I have been mulling on for a while. As I used it, I began seeing how the argument of these generators being tools hold ground, having experienced using them to flesh out an idea, through words, can not only deliver a result, but deliver a result that I’d never probably never have thought of – Pascale Blanché has some outstanding explorations on his twitter feed. Seeing how this workflow can push one’s creativity in new directions is quite astonishing, especially when we consider all artists draw on internal and ‘physical’ libraries of reference, coupled together with their own thought processes, thus can be prone to subconsciously rehashing the same ideas or style.

The above image was created through generating multiple images, with different prompts, and then compositing, editing and over painting them until I reached the idea I was building on. While the base images are generated by the AI engine (guided by my prompts and decisions on refinement direction), the final image is very much my own, based on my ‘manual’ input.

Is it ‘my‘ art? I would have to say yes – the raw images are very different from the final. But it’s art that’s arrived at a final point via a very different road we so far have been accustomed to. The question I am now asking myself is if these AI art tools are any different from the myriad of methods, such as ‘image bashing’, compositing, referencing etc., already used by professional creatives as part of their daily workflow?

 

0

Copyright 2022 Gerard Thomas. All rights reserved.