self-portrait* in blue

August 13, 2022

^{[I originally posted this entry as early access for my Patreon patrons on 7 August 2022].}

I had two other prompts already written down to try with Midjourney AI before deciding it would be remiss not to try something (in theory) more personal.

I was curious how much - if at all - the AI was trawling the internet for references.

I'd seen people entering 'in the style of [artist name]', and I'd obviously used 'art deco' in my previous prompt.

So, my mischievous side came out and for my third and final prompt before my trial ran out, I entered 'bronwen hyde, self-portrait, blue'.

The initial four options and the second four options Midjourney gave me can be seen in the second image.

I wavered a little between the image I chose and the top right and bottom left options in the first four. The second set of four didn't appeal to me as much stylistically.

In the end, as I had only a few variations and upscale "turns" left, I worked just with the top left image from the first four options. This was my final result before my trial ended.

Clearly, it's not me, but I feel like some options remind me of Cindy Sherman and maybe Sigourney Weaver. So it is interesting to think about how Midjourney learns and what it picks up as sources.

I'll probably sign up for a subscription for a month to play with the tool more and see whether it's something I find helpful as a part of my art practice.

But either way, it was an engaging and intriguing way to spend a couple of hours on a Friday evening.

art deco annie

August 13, 2022

^{[I originally posted this entry as early access for my Patreon patrons on 6 August 2022].}

The second prompt I gave Midjourney AI was 'art deco, portrait, night, lamp'.

This image was my favourite result, though I have to admit I didn't play too much with creating variations for this. I felt with a few tweaks, it came up well.

I did get the AI to give me an alternative set of options before I received a notification that I had a limited number of free images left to create.

In the second image (left to right), you can see the original four options Midjourney generated for my prompt. And then the second four options.

death in the afternoon

August 13, 2022

^{[I originally posted this entry as early access for my Patreon patrons on 6 August 2022].}

Yesterday afternoon, scrolling through my feed on Facebook, I saw some artwork by a fellow photographic artist I follow, Rebecca Tolk, and was intrigued by it. Especially when she mentioned the work was created using artificial intelligence.

Although the work looked amazing, my initial thinking was similar to what I soon found out was Rebecca's initial question: "How is this her art?"

Wanting to understand more, I watched the replay of a Facebook Live video she made. Rebecca explains how she sees this tool fitting into her practice and how she has seen others use it in their practice.

If you have the time, it's well worth a watch to understand the 'why', and it gives a sense of how I could potentially see it fitting into my practice too.

For example:

To create elements for inclusion in mixed media artworks without having to go out and photograph each of them (especially if they're not easily accessible for me).
To create elements for my digital collages.
To inspire and draft out ideas for photographs or collages.
To storyboard potential short films I'd like to make, especially to give a sense of mood and atmosphere. This would be particularly helpful as my illustration skills are still quite rudimentary.

And, obviously, there's the option to use it in the most straightforward way to create art in and of itself. Though I think I would have to become more familiar with how it works and how to "work it" to consider it more "my artwork" than the AI's.

So, I took some time yesterday afternoon to experiment with a free trial of Midjourney AI to see what I thought.

As Rebecca points out in her video, your free trial images and variations don't take long to exhaust. And working in a massive group chat with so many other people can be overwhelming. Though also inspiring to see what other people are using as prompts and how they are using Midjourney.

It can be hard to track your results, and I'm not entirely sure I found all my variations as I was still learning to use the tools. If I decide to sign up for a limited subscription, I believe you get your own workspace. That will make it easier and cleaner to focus on what I'm doing and learn more about how to use it.

But these are two variations from the prompt 'death in the afternoon' that I created yesterday afternoon.

The first image was actually my second option and variation, but it was the one I liked more. The branch coming out of nowhere is the sort of thing I would edit out as it's obviously a bit of a glitch in the AI. In the preview, it looked like it might be the blade of a sickle.

The third image shows (from left to right):

The initial selection of four AI-rendered images Midjourney presented me with, of which I selected the top left.
The variations it created from that image when I triggered it, of which I selected the bottom right.
An alternative set of four AI-rendered images Midjourney presented when I requested a redo. I didn't get to work with these as I didn't see them initially, and I used up the rest of my free trial working on two other prompts.

I definitely don't see AI art replacing my existing practice. But I can see how it might enhance my work and/or be another tool or medium for me to create art.

I'll share the other two images I created yesterday later in the day, but I'd be curious to know your thoughts.