Using Midjourney
The AI workshop allowed me to explore new ways and foundations to start making concepts from any ideas that I have. I used mid-journey’s discord server to create prompts that generated AI images based on the description.
I found this result of the prompt to actually meet my expectations of what I was expecting it
to look like. It could be said that the AI used MF Doom’s abstract style of hip hop and lyrics to rejuvenate that same abstract style but in a painting rather than a realistic image.
Overall I found the AI very successful in interpreting my idea. It was able to actually create a concept that looked somewhat close to what MF Doom would look like running. However I will say that the one weakness the AI had was getting legs to look right on the 2nd image
From here, because I was impressed, I wanted to see if the AI on mid journey could generate a more complex prompt. The prompt was “Spike Spiegel fighting Goku.” This one didn’t really work and instead of showing them fighting each other, it just merged the 2 characters together.
However one thing I did come to realise at this point is that in using the prompt for AI to generate your concept, you have to be more detailed and specific so that it can get a more accurate result of what you want. For example, instead of saying ‘Spike Spiegel fighting Goku,’ I could’ve thrown in “each other” in the prompt so that it may have not merged the two characters together. However one thing that I noticed throughout these workshops is that AI is overall accurate on obtaining and exhibiting the right colour schemes based on your prompt and the mood you ask for.
Using Dall-E
Dall-E is a machine learning model developed by OpenAi that generates digital images from your prompt descriptions. I found that using Dall-E was able to create a more accurate image from a more complex prompt I was using.
As an example my favourite outcome from using this site was the prompt Takashi Murakami style painting of God. I didn’t know what to expect to see from this but the result was interesting yet not shocking. Interesting because its depiction of God reflects a Buddhism image of what that would look like which isn’t suprising because Takashi Murakami is from Japan where that religion is mainly believed in. I think it was quite accurate in getting the colour scheme and abstract style of painting right.
Finally to conclude the workshop, I wanted to show how I can use AI to create or finish my own concepts for my own work by me. I used one of my paintings and asked 2 people what 3 words they felt from my artwork.
Person 2: spirituality, love and doubt.
Person 1: authentic, meaningful, thought evoking.FKA Twigs. Nujabes
From these adjectives I was able to have enough words to try and use Dall-E to finish this piece of artwork, or rather expand upon it.
After many prompts and variations later I had this as the final piece and what came to show was a part AI part handcrafted artwork. I thought that the AI was very accurate in creating the abstract style of the artists Nujabes and FKA Twigs on the left side of the painting and also picking up on the correct colour tones that I used. However, one thing I did feel is that I preferred the painting unfinished or in my case how it originally looked. This isn’t to say that what the AI did was good, but I think it is a case where you either love it or hate it when AI is developing your own work depending on what you’re looking for.
On concluding this workshop my views on AI have changed. It went from being skeptical and discouraging for using it as an artist to using it as a way to create quick and rough concepts for your own thoughts and imagination. The art world is always constantly changing and is adapting faster than ever. I feel as if AI is here to be interpreted by artists for a reason whether we like it or not, so using it should be a tool that we could least be open to using. Even if it is just to simply create concept art or use a prompt to generate some ideas for backgrounds, colour or even character ideas for our own projects, I think I will somehow see if I can use it for my own final major project for developing ideas in the concept art and storyboard part in my production. Or another interesting and innovative approach for using AI is to let it finish and generate abstract images of artwork that we have already done along the line of our production for the project and see if it makes something look better that we can include in the final outcome.