As not enough people have noticed..
AI has made huge progress when it comes to art. A year ago AI-images were obviously computer-made; they looked off and derpy, every now and then bordering terrifying. And while there are still some things the AI struggles with (hands), most images that come out of an AI like MidJourney will fool the general and specialised public as to how it was made. The results its generating are so beautiful that every creative should feel threatened as well as intrigued. After discovering MidJourney I myself fell in a deep AI-bordered hole I need to talk about, so here we go.
First off, the process of generating art with an AI is stupidly simple. You go to the AI generator’s website, give it a prompt and it will make whatever you ask it. For the image below I used “Jazelle uglyworldwide photographed by nick knight, a new fashion series, old english house, soft light, 35mm, canon, hasselblad” which references two amazing current artists (which is very much frowned upon, more on that later), a lot of photography terms and the light and setting I wanted to achieve. In a minute you get four sketches, which you can upscale to usable images afterwards (on the right):
It might take a couple of tries, but once you get the prompts down you can create whatever is in your mind (also more on this later). For the now-enthusiastic-creative: I use MidJourney. You can join for free and get about 25 tries for free. After that it’s about €12 per month for 200 images. I’m well on my way through those after a week. Still, totally worth it to me.
Some problems with this
And considering MidJourney has been online for a short period of time, since march 2022, I think we can safely say that photographers and other creatives are going to feel the heat next. I envision so much customized stockphoto’s, deepfakes and in general, AI-photography to flood the market once this gets some traction. One of the MidJourney-options is to feed photo’s into the prompt, so in a year you can probably add your selfie and ask the AI to make a LinkedIn suitable image for you without a problem. Or add this image of model X plus a photo of the outfit from your brand to make a more beautiful campaign than you would normally have money for. This will cause some copyright issues, which I don’t think will even be the real problem. (Interesting to note is that MidJourney states that the he AI only references images, never uses them completely.)
The real problem
If you think about it, AI might just turn out to be a tool like photography was in the beginning – and look where we are now. Everybody likes to take pictures but there’s still a market for specialists, people that are better at it than others. There’s more interest in and more money spent on art than ever before; never was it valued so widely as it is today. AI images will kill all of our conceptions of art, but I think the next step might be the logical next step if you look at modern art anyway: the idea ís the art.
Images all inspired by text from “The future is faster than you think”.
The future - how?
Quite possibly there wil be a huge future for creatives in the field of AI. I think we need artists to be visionaries, philosophers and thinkers to create this field. What images and artist you can reference, what taste you have, which artistic choices you make will be the most important tool for any artist. And that has not been that different lately. Especially in photography you see crazy amounts of repetition as it has been hard to still have an original, creative voice. As I said; what even is originality? Other than taking existing ideas, combining and building on those? True originality is hard to come by and will be more and more scarce – something you can see happening on instagram/tiktok, and definitely on the discords of MidJourney, where a lot of the creators are churning out very similar images drawing on cyberpunk, hot people, fantasy and/or whatever they’re a fan -or hater- of.
To know your own style, likes and dislikes will be more important than ever. Some creatives that are pioneering the field are Sentientmuppetfactory, Holly Herndon, Nick Knight himself with his NFT’s (which already feels old, but still) – and within all you can detect their own distinctive style. While it may not be groundbreaking to use the 70’s as a nostalgic reference, creator Beth of Sentientmuppetfactory still finds a way to really connect these very odd images to something that obviously brings feelings to a lot of people. It’s funny, scary, odd yet recognisable… And she painstakingly creates something we haven’t seen before.
Some more issues
And to get back to the issue of ethics; one of the things the current MidJourney community frowns upon is the usage of currently-alive-artists to create your own prompts. It invades their livelihood, their artistic voice and can therefore disrupt their own practice. While I personally think a rule like this feels “decent”, there are still a lot of copyright questions here. How influenced can you be by someone’s style without infringing their copyright? I saw someone create a perfect Disney catalog page (Disney being notorious for copyright charges) on the discord – using that won’t go down well. But considering everyone and their mom can use AI now, is there still use for these copyright claims?
About the images I made
In my images here I used the photographer Nick Knight and model UglyWorldWide as a prompt, who to me are the epitome of creative artists. But I do feel these images are more their work than my own. Making these I realised my own style is still something different, maybe a bit more soft-spoken and visually not as strong as his yet. It was a nice reflection to use – and to build upon – in my own work. I would probably use these (as I am using them in this article) but I would also reference the artists that influenced this. And to be honest: It feels rather debatable whether creating images with other people’s work as a prompt is completely ethical.
To see how aware others would be I (being a photographer) tried posting an AI-made fashion series to my Instagram. All I got were “wow”’s and zero “this looks odd or fake”’s as people never noticed I didn’t pick up my camera for this. The people who follow me are usually more involved in art or photography then the general public – that’s why they follow me. So it is a bit concerning that these “specialists” didn’t notice a thing. I must admit the work I put out was hauntingly beautiful, above my skill set I would say even. I realized my ego hurt a little when it lost the battle to an AI when it came to creating beautiful imagery.
Images all made by me and MidJourney, and inspired by Makoto Shinkai, Lobsters and fashion, 70’s posters and an Iranian elder lady photographed by Nick Knight.
On the topic of loss aversion
And as I was trying to talk to other creatives about the subject of AI made art, I noticed it was not something people wanted to think or talk about. And oftentimes people tell me this is “not art” and I shouldn’t post it as “I didn’t create it”. While I think there is some truth in this (I did not pick up a camera), I also feel its denial of new potential. Drawing or even visualizing has never been my strong suit – but I am a damn creative person with a huge mind-library for references. I can combine things others wouldn’t think of and I can now, more than ever, create those things oh-so-easily. I could fight this coming wave of AI-art — but I’d rather learn to surf it. So surf with me, creative people, and let’s talk about in which direction we want to head.
Email me if you’re interested! 💌
Maan
P.S. Interesting reads / watches:
Book: “The future is faster than you think – How Converging Technologies Are Transforming Business, Industries, and Our Lives” by Peter H. Diamandis & Steven Kotler (all the disruptive technologies that are coming in one book, go read it)
Video: “The text-to-image revolution, explained” by Vox (Thank you for the tip, Jean-Sébastien Monzani)
Video: “Who Is Midge? Midjourney Slang, Discord Tips & Other Tidbits | Midjourney Tutorial” By Future Tech Pilot (on Midjourney slang)
