I remember when photoshop became widely available and the art community collectively declared it the death of art. To put the techniques of master artists in the hand of anyone who can use a mouse would put the painter out of business. I watched as the news fumed and fired over delinquents photoshopping celebrity nudes, declaring that we’ll never be able to trust a photo again. I saw the cynical ire of views as the same news shopped magazine images for the vanity of their guests and the support of their political views. Now, the dust long settled, photoshop is taught in schools and used by designer globally. Photo manipulation is so prevalent that you probably don’t realize your phone camera is preprogrammed to cover your zits and remove your loose hairs. It’s a feature you have to actively turn off. The masters of their craft are still masters, the need for a painted canvas never went away. We laugh at obvious shop jobs in the news, and even our out of touch representatives know when am image is fake.
The world, as it seems, has enough room for a new tool. As it did again with digital photography, the death of the real photographers. As it did with 3D printing, the death of the real sculptors and carvers. As it did with synth music, the death of the real musician. When the dust settles on AI, the artist will be there to load their portfolio into the trainer and prompt out a dozen raw ideas before picking the composition they feel is right and shaping it anew. The craft will not die. The world will hate the next advancement, and the cycle will repeat.
When it comes to AI art, the Photoshop/invention of the camera argument doesn’t really compare because there’s really 2 or 3 things people are actually upset about, and it’s not the tool itself. It’s the way the data is sourced, the people who are using it/what they’re using it for, and the lack of meaning behind the art.
As somebody said elsewhere in here, sampling for music is done from pre-made content explicitly for use as samples or used under license. AI art generators do neither. They fill their data sets with art used without permission and no licensing, and given the right prompting, you can get them to spit out that data verbatim.
This compounds into the next issue, the people using it, and more specifically, how those people are using it. If it was being used as a tool to help make the creation process more efficient or easier, that would be one thing. But it’s largely being used by people to replace the artist and people who think that being able to prompt an image and use it unedited makes them just as good an artist as anybody working by hand, stylus, etc. They’re “idea” guys, who care nothing for the process and only the output (and how much that output is gonna cost). But anybody can be an “idea” guy, it’s the work and knowledge that makes the difference between having an idea for a game and releasing a game on Steam. To the creative, creating art (regardless of the kind - music, painting, stories, whatever) is as much about the work as it is the final piece. It’s how they process life, the same as dreaming at night. AI bros are the middle managers of the art world - taking credit for the work of others while thinking that their input is the most important part.
And for the last point, as Adam Savage said on why he doesn’t like AI art (besides the late-stage capitalism bubble of it putting people out of work), “They lack, I think they lack a point of view. I think that’s my issue with all the AI generated art that I can see is…the only reason I’m interested in looking at something that got made is because that thing that got made was made with a point of view. The thing itself is not as interesting to me as the mind and heart behind the thing and I have yet to see in AI…I have yet to smell what smells like a point of view.” He later goes on to talk about how at some point a student film will come out that does something really cool with AI (and then Hollywood will copy it into the ground until it’s stale and boring). But we are not at that point yet. AI art is just Content. In the same way that corporate music is Content. Shallow and vapid and meaningless. Like having a machine that spits out elevator music. It may be very well done elevator music on a technical level, but it’s still just elevator music. You can take that elevator music and do something cool with it (like Vaporwave), but on its own, it exists merely for the sake of existing. It doesn’t tell a story or make a statement. It doesn’t have any context.
To quote Bennett Foddy in one of the most rage inducing games of the past decade, “For years now, people have been predicting that games would soon be made out of prefabricated objects, bought in a store and assembled into a world. And for the most part that hasn’t happened, because the objects in the store are trash. I don’t mean that they look bad or that they’re badly made, although a lot of them are - I mean that they’re trash in the way that food becomes trash as soon as you put it in a sink. Things are made to be consumed in a certain context, and once the moment is gone, they transform into garbage. In the context of technology, those moments pass by in seconds. Over time, we’ve poured more and more refuse into this vast digital landfill that we call the internet. It now vastly outweighs the things that are fresh, untainted and unused. When everything around us is cultural trash, trash becomes the new medium, the lingua franca of the digital age. You could build culture out of trash, but only trash culture. B-games, B-movies, B-music, B-philosophy.”
That is precisely it. Generative AI is a tool, just like a digital canvas over a physical canvas, just like a canvas over a cave wall. As it has always been, the ones best prepared to adapt to this new tool are the artists. Instead of fighting the tool, we need to learn how to best use it. No AI, short of a true General Intelligence, will ever be able to make the decisions inherent to illustration, but it can get you close enough to the final vision so as to skip the labor intensive part.
On this topic, I am optimistic on how generative AI has made us collectively more negative to shallow content. Be it lazy copypaste journalism with some phrases swapped or school testing schemes based on regurgitating facts rather than understanding, none of which have value and both of which displace work with value, we have basically tolerated it.
But now that a rock with some current run through it can pass those tests and do that journalism, we are demanding better.
Fingers crossed it causes some positive in the mess.
Not the same thing, dog. Being inspired by other things is different than plagiarism.
Humans learn from other creative works, just like AI. AI can generate original content too if asked.
AI creates output from a stochastic model of its’ training data. That’s not a creative process.
What does that mean, and isn’t that still something people can employ for their creative process?
A person sees a piece of art and is inspired. They understand what they see, be it a rose bush to paint or a story beat to work on. This inspiration leads to actual decisions being made with a conscious aim to create art.
An AI, on the other hand, sees a rose bush and adds it to its rose bush catalog, reads a story beat and adds to to its story database. These databases are then shuffled and things are picked out, with no mind involved whatsoever.
A person knows why a rose bush is beautiful, and internalises that thought to create art. They know why a story beat is moving, and can draw out emotional connections. An AI can’t do either of these.
LLMs analyse their inputs and create a stochastic model (i.e.: a guess of how randomness is distributed in a domain) of which word comes next.
Yes, it can help in a creative process, but so can literal noise. It can’t “be creative” in itself.
How that preclude these models from being creative? Randomness within rules can be pretty creative. All life on earth is the result of selection on random mutations. Its output is way more structured and coherent than random noise. That’s not a good comparison at all.
Either way, generative tools are a great way for the people using to create with, no model has to be creative on its own.
How that preclude these models from being creative?
They lack intentionality, simple as that.
Either way, generative tools are a great way for the people using to create with, no model has to be creative on its own.
Yup, my original point still stands.
How is intentionality integral to creativity?
LLM AI doesn’t learn. It doesn’t conceptualise. It mimics, iterates and loops. AI cannot generate original content with LLM approaches.
Interesting take on LLMs, how are you so sure about that?
I mean I get it, current image gen models seem clearly uncreative, but at least the unrestricted versions of Bing Chat/ChatGPT leave some room for the possibility of creativity/general intelligence in future sufficiently large LLMs, at least to me.
So the question (again: to me) is not only “will LLM scale to (human level) general intelligence”, but also “will we find something better than RLHF/LLMs/etc. before?”.
I’m not sure on either, but asses roughly a 2/3 probability to the first and given the first event and AGI in reach in the next 8 years a comparatively small chance for the second event.
