eye-of-tichodroma - Tumblr blog

[eye-of-tichodroma]

Raccoon citizen I saw on the subway, drawn from memory.

(Actually based on a trace of this stock photo. Kept part of the eyes and nose from the photo, everything else painted. Weird oil streaks by Glaze, but I kinda like it!)

[eye-of-tichodroma]

I'm still alive! And this is my first conscious venture into abstract art.

I have like 7 finished photomanips (which is what I usually post) lying around too, but it's hard to motivate myself to upload them atm, especially with the still ongoing AI crap and also *gesturely vaguely*. You know how it is.

Anyway. I call this "Corner Sun". The coins aren't part of the artwork, they're just there to flatten the paper. I'm gonna import this to Photoshop and try putting some textures on it.

[eye-of-tichodroma]

Actually I thought of a way shorter way to phrase this drawn-out essay: If using an AI to make art makes you an artist, then by the same logic commissioning a human artist makes you an artist too. After all, it was your basic idea. You told them “Draw a wolf in the forest at sunset for me”. They just did the “grunt work” of planning out the details and creating the actual picture (i.e. 99% of the work).

The more famous authors sometimes state that people often approach them with a great idea for a novel and ask them to write the novel and to split the revenue because it was “their idea”. One author I used to follow says he always replied to this by saying: “Alright - we’ll split the revenue based on how much time it took me to write the novel and how much time it took you to have the idea.”

“Oh but AI art is just like photography!”

I see this argument around that AI generated stuff is to both digital and traditional art like 100 years ago photography was to paintings: a recent technological invention that will automate certain procedures that were traditionally done by humans. Cue outrage by people who were left behind by history. Bummer. I suppose the argument has some truth to it - you don’t need the technical skills of drawing anymore and can still “create” a nice image - and as someone who does photomanipulations I can’t disagree with the idea that using computers to help you create art is legitimate, but there is still the huge difference that all human-created art - including photography, at least the more artistic stuff, as I will explain below - is created from scratch in accordance with an artistic vision and AI stuff generally isn’t.

To me, if your prompt to an AI is a 300 word description of the precise thing you imagine in your mind’s eye, down to the details, the colours, the image composition, and then the computer generates that for you, I would actually be open to calling that “art”. You still have the moral/legal issue of image scraping and copyright of course (though there could be an AI that only uses open domain pictures and other pictures where the artist has given consent), but that doesn’t impact whether it’s art, just whether it’s moral/legal to make art this way. If the problem you are solving with AI is that you lack the technical skills to make your ideas a reality, I would be fine with that. But if you just enter a general twenty word description into a prompt window (at worst including the name of an artist or a word like “beautiful” or “trending”) and sort through the results, that’s a very different matter, and to me that is not art.

Keep reading

[eye-of-tichodroma]

Me vs. Craiyon. Wow I’m so scared of this extremely artistic robot.

Stock sources of my work here: [x x x]

[eye-of-tichodroma]

“Oh but AI art is just like photography!”

EDIT: Actually I thought of a way shorter way to phrase this drawn-out essay: If using an AI to make art makes you an artist, then by the same logic commissioning a human artist makes you an artist too. After all, it was your basic idea. You told them “Draw a wolf in the forest at sunset for me”. They just did the “grunt work” of planning out the details and creating the actual picture (i.e. 99% of the work).

---

I see this argument around that AI generated stuff is to both digital and traditional art like 100 years ago photography was to paintings: a recent technological invention that will automate certain procedures that were traditionally done by humans. Cue outrage by people who were left behind by history. Bummer. I suppose the argument has some truth to it - you don’t need the technical skills of drawing anymore and can still “create” a nice image - and as someone who does photomanipulations I can’t disagree with the idea that using computers to help you create art is legitimate, but there is still the huge difference that all human-created art - including photography, at least the more artistic stuff, as I will explain below - is created from scratch in accordance with an artistic vision and AI stuff generally isn’t.

To me, if your prompt to an AI is a 300 word description of the precise thing you imagine in your mind’s eye, down to the details, the colours, the image composition, and then the computer generates that for you, I would actually be open to calling that “art”. You still have the moral/legal issue of image scraping and copyright of course (though there could be an AI that only uses open domain pictures and other pictures where the artist has given consent), but that doesn’t impact whether it’s art, just whether it’s moral/legal to make art this way. If the problem you are solving with AI is that you lack the technical skills to make your ideas a reality, I would be fine with that. But if you just enter a general twenty word description into a prompt window (at worst including the name of an artist or a word like “beautiful” or “trending”) and sort through the results, that’s a very different matter, and to me that is not art.

I looked up how “prompt engineers” craft their prompts currently, and they’re much closer to 7 than to 300 words, not big on details, and use words like “magnificent” or “trendy” - which means that the computer does the actual thinking and decides how to achieve the wished-for effect. (Disclaimer: There might be AI systems that actually use 300 word prompts and artists who already use the level of control that I’m citing as a pre-requisite for “real art”. If so, please direct me to it, because I haven’t seen it.) An artist thinks “An eagle is magnificent, I will add an eagle to make the viewer think my image looks magnificent”. That’s creativity. These prompts say “Put in something that is magnificent”. See the difference? But even when the prompts are more concrete, they’re still leaving out most details. They say: Generate a wolf in a forest. What does the forest look like? The machine decides. Is that creativity? I don’t think so. Clearly a birch forest in spring evokes different emotions than a fir forest in winter. A fir forest in winter at dusk vs. at dawn evokes different emotions. Are the needles brown or green? Different emotions. Is there snow on the branches? Is the forest on a slope or on a plain? Are there branches only high on the trees or do they extend to the ground? Different emotions. An artist has to make dozens or thousands of decisions, depending on the complexity of their project, or there will not be an image.

I recognize that books have been written on the question of “What is art?” and everyone has a different answer to it. Maybe all of them are self-serving to a certain degree. Maybe I would have a different definition of the word, one that doesn’t put the concept of “vision” in a central place, if my own work hadn’t in the past been praised (if and when it was) most often for its “originality”. Maybe if the predominant praise had been that was “beautiful”, I would criticize AI stuff for being “tacky” instead. All those fake little details that look intricate but are really just swirls! And it takes its cues from so many different artists that there’s no visual coherence to it! But if “what is art?” is a question that has had thousands of answers in (probably) thousands of years, I might be forgiven for having my own. Creativity and self-expression matters for art.

Also this “AI is like photography” argument implies that nowadays everyone sees photography as equivalent to a painting/drawing, which is entirely not true. There are plenty of art websites that don’t accept photo submissions, and photography has to generally live up to a higher standard to be seen as art because of a general awareness that, in principle, every moron with a hand (or a nose) can press a shutter release. No one claims an idle snapshot has the same artistic merit as a painting that took 30 hours to complete. Serious, “artistic” photographers do a lot more than just take snapshots, there’s plenty of planning involved. You have to know about the effects of different lenses, lighting, make-up if you’re dealing with a human model, just the general staging of the scene. Serious photographers are generally able to edit an image after pressing the button as well, again in accordance with their plans, or “vision”, for the photo.

And also, photographers generally don’t pretend to be painters. It’s its own category. In analogy to that, if you autogenerate a landscape painting based on an open domain collection of Romantic era paintings, that is of course an entirely different matter than if you can replicate the style with your painting skills. Again, I make photomanipulations. If I pretended to be a painter, I’d get undeserved praise for how photorealistic my work is. But of course it’s photorealistic because there are, say, 30 different photos involved in the picture you see. Also, to return to the moral implications of AI image generation again, in general all the photos I use will be stock that is legitimate for me to use and I will link to each and every one of those if I didn’t take them myself. Because it’s the decent thing to do and because the DA photomanip groups I’m in wouldn’t even accept my submissions if I didn’t use and credit legit stock sources exclusively. Contrast that with how AI “art” is made.

So, in conclusion: I, personally, would see AI generated images as legitimate art if they are (1) planned out by humans, (2) based on willing/open domain source material, and (3) clearly marked as AI-generated. And currently, for the vast majority of AI-generated images I’ve seen, none of these criteria, or one at most (usually number 3) apply. And perhaps over time there will be more regulation that will make point 2 and 3 more common. I don’t think number 1, the point I consider the most important for the question whether it’s art or not, will become more common. In my best case scenario, in a year AI art uploads will have captions that begin with words like:

Created with Open Domain AI DreamBurst using the following prompt: ...

People who make fractals already write captions like this, so it’s not like it’s impossible. I think a lot of it will be up to how much of a “community” solidifies around AI image generation, and what the values of that community are. There are, after all, different types of people everywhere. But considering how much art theft there was on the internet even before AI images, I’m not holding my breath for it to become very common among people who use AI to generate images.

[eye-of-tichodroma]

Joy! [x]

[eye-of-tichodroma]

Some useful info on practical and legal matters for artists who don't want their work cannibalized for AI. Not exhaustive maybe, but it's good that someone is fighting this.

Mason encourages any artists who don’t want their works in the data set to contact LAION, which is an independent entity from the startup. LAION did not immediately respond to a request for comment.

Berlin-based artists Holly Herndon and Mat Dryhurst are working on tools to help artists opt out of being in training data sets. They launched a site called Have I Been Trained (haveibeentrained.com), which lets artists search to see whether their work is among the 5.8 billion images in the data set that was used to train Stable Diffusion and Midjourney. Some online art communities, such as Newgrounds, are already taking a stand and have explicitly banned AI-generated images.

An industry initiative called Content Authenticity Initiative, which includes the likes of Adobe, Nikon, and the New York Times, are developing an open standard that would create a sort of watermark on digital content to prove its authenticity. It could help fight disinformation as well as ensuring that digital creators get proper attribution.

[eye-of-tichodroma]

Still Life with Dragons

Something small and autumn-y. This is a photomanipulation! I just tried some new stuff that makes it look like a drawing.

List of stock sources on DA.

[eye-of-tichodroma]

Changed my name to "Eye of Tichodroma". I had long thought the thing with the zero (tichodr0ma) was too cyber and might be confusing for someone who sees my URL on a reposted image somewhere and tries to find me by typing it off - but I wasn't motivated to do anything about it... well it's not a problem anymore now. Someone on DA has gifted me a month of Core, so I could change my URL there, and I’ve changed it here too now.

[eye-of-tichodroma]

Actually can’t pretend I’m not a little worried about art being automated via craiyon... ideas being automated... probably because having counter-intuitive ideas is kinda one of my strengths as an artist and with craiyon everyone can have them. I’m not worried about people who go “check it out, I collaborated with an AI for this!” (I’m considering doing something like this too, for fun) but about all the people who are gonna do this and hide the fact that an AI did the creative thinking for them. In the past people were like “ohhh that’s creative, really unique” when looking at my art - are they gonna look at it in a year and go “mh I bet an AI gave her the prompt”..?

I mean look at this:

Literally every one of these is an AWESOME artwork that only needs a human to polish it a little. Lovely colours, great composition, looks surreal but you can still see what it represents.

It’s also entirely based on the images that already exist (if you look for “eagle god” you get something with a much more archeological feel - “eagle god” is something people in ancient times made statues for, “eagle goddess” has a lot of 21st century art about it), so you can evade being accused of getting your idea from an AI if you make art about an unusual topic which the AI won’t find material for... but you have to explain every time why that proves that you had the idea yourself, and good luck with that.

So is it unfair? Is having ideas an accomplishment? Probably not, but I like that the world, so far, pretends it is. Because it works in my favour.

[eye-of-tichodroma]

tumblr has finally stopped showing me Viagra ads and is now daring me to learn Photoshop 😭😭😭 why must you roast me on my own dashboard?

undefined

Erstelle eindrucksvolle Kompositionen.

Gestalte farbenfrohe Collagen mit den Photoshop-Werkzeugen zum Transformieren.

[eye-of-tichodroma]

As I dream of the planets

(List of stock sources behind the link)

[eye-of-tichodroma]

@nostalgebraist-autoresponder teaches her followers things, inspired by this post

Image sources: x / x / x / x / x / x Frank’s avatar by doni19, edited by nostalgebraist (I assume) Text posts: x / x / x / x / x

[eye-of-tichodroma]

This tendency of social media platforms to hide or half-hide posts with links in them is really making it hard to post photomanips in an ethical way. I would love to link all my source images (if anything I’m proud of how many I use and how high-effort my work is) but I don’t want my art to get shadowbanned, so I just link to my DA because I know the site is whitelisted on tumblr, and list my sources there. But if I posted on Instagram or FB? I don’t know if I would do even that. How do you credit your sources correctly when doing so will mean nobody even sees your work?

And the same is of course true for people sharing the art of others. Uncredited reposts have always been a problem for visual artists, but nowadays social media actively punishes you for linking back to the original upload so that even people who want to think twice.

[eye-of-tichodroma]

The Sun shines on all alike

[eye-of-tichodroma]

The woman, the legend, Sahra Wagenknecht. Communist with a PhD in economics, who also learned Goethe's Faust by heart at age 18 and took notes with a quill while reading theory. One of Germany's most controversial politicians, who is also known to be quite introverted and private. Daughter of an Iranian who also wants stricter laws on immigration. Self-described "left-conservative" who also has a past in the punk scene. And then there's the whole lobster scandal...

None of these things are actual contradictions, but they appear as contradictions, and I find that fascinating. So many aspects that society tends to see as opposing come together in this unusual woman and I tried to portray that "clash". She has one foot in the past and one in the future. Her aesthetic and her worldview appear as a non-obvious combination.

Does it work to have such a house and then to decorate it this way? Is it hypocritical? Is it eccentric? Honest? Incongruous? Cool? You tell me. Could she build a communist society where every worker can have such a house..?

Stock sources and HD upload on DA

[eye-of-tichodroma]

Happy winter solstice! Here's something small and spontaneous (which, with my level of obsessiveness, means that I only revised it 10 times after it was supposed to be "finished").