A friend asked, "Is it an overreaction on my part to unfriend people who publish and support A.I. generated images...?"
Let me say unequivocally and emphatically that the rise of A.I. image making systems is an ASSAULT ON THE ARTISTIC COMMUNITY in every possible respect; an illegal, unethical and amoral pillage and plunder offensive, enabled by a perfect storm of greed, ignorance and entitlement. Shall I go on? Ok, I will.
Though I believe a lot of the folks "using" (being used by) A.I. image generators are unaware of the potential consequences, the companies behind them are FULLY aware of what they are doing, and it is 100% intentional. Their goal is to replace, not empower human artists.
Can we not reserve the most precious aspects of civilization for HUMANS? The capitalist, consumer-driven marketplace says emphatically: NO. For that thing it’s all about how to generate the biggest pile of good enough, “consumable content” with the least expense and risk. This is exactly the opposite of creativity, artmaking, and culture.
Three relevant points for me: 1. Current generative A.I. systems are illegally exploiting artists’ work 2. A.I. imagery is not made "USING” artificial intelligence—it’s made BY artificial intelligence. 3. That distinction matters a lot
1. THE LEGAL ISSUES
Here’s why Stable Diffusion says they only use copyright-free MUSIC for Dance Diffusion:
“Because diffusion models are prone to memorization and overfitting, releasing a model trained on copyrighted data could potentially result in legal issues. In honoring the intellectual property of artists while also complying to the best of their ability with the often strict copyright standards* of the music industry, keeping any kind of copyrighted** material out of training data was a must.”
Translation: We violated the copyright of millions of visual artists because, unlike with the music industry, we thought we could get away with it.
Notes: * The “standard” of copyright is just as firm for works of visual art, writing, etc.—it’s just that the music industry has historically ENFORCED copyright more effectively;
** Material is not “copyrighted.” Copyright is noun, not a verb. It's a RIGHT held by the author of a work. If you made it, only you can make copies of it, because only you have copyright (there are well intentioned exceptions, which A.I. systems have exploited in bad faith)
2. WHO OR WHAT MAKES AN A.I. GENERATED PICTURE?
When Midjourney produces an image, where is that image mostly coming from? The A.I. user’s prompting is nigh insignificant compared to the billions of hours of work put in by artists whose art was misappropriated to create the training dataset, not to mention the work being done by the A.I. itself.
But because there are no other hands or eyeballs in sight, the prompter claims the title of artist or creator by default. “It’s my idea, so I must be the creator!” This is convenient if pathological. An art director giving the idea, instructions and ongoing input for a work of art is not the artist; an editor giving assignments and editing stories is not the writer.
A.I. defenders continue to make cherry picked, superficial comparisons to previous technologies such as Photoshop. But actual artist TOOLS like these are simply digital corollaries of their traditional (physical) predecessors. That’s why they employ “brushes” and “erasers” and “palettes.”
By contrast, an A.I. image generating system IS A COROLLARY OF A HUMAN ARTIST. It is an “intelligence” that uses “prompts” (INSTRUCTIONS), then you wait while it makes pictures. If it took the form of an android holding a paintbrush, and required a couple of days to generate an image, maybe its nature would be more obvious? Some liken it to the camera, because "You merely press a button and get a picture." But the camera is a corollary of the EYE not a brain.
So where is the example of a previous technology to which you give verbal instructions, then it creates a new picture for you? There isn’t one, or at least, we don’t call it technology—we call them “artists.”
Let me put this very bluntly: MIDJOURNEY IS SMARTER THAN MOST OF THE PEOPLE WHO USE IT. It has a higher level of artistic intelligence, which is why they need it.
3. WHY IT MATTERS
A.I. users say “if the picture is good it won’t matter where it came from.” I don’t believe that. Consider a child’s heartfelt drawing of her world vs. A.I. mimicry that creates an identical image. If we’re told a child made it, we’re touched; if we’re told an A.I. made it, we’re uninterested or even put off. THIS IS NOT A FLAW IN OUR WIRING. Cave paintings with the artists' handprints are mind blowing, BECAUSE HUMANS LIKE US ACTUALLY MADE THEM. Art connects humans to one another through direct transmission of lived experience.
Someone invented a device that can make pictures better than most people can, and better than a lot of people who want to call themselves artists can. Big whoop. We have vehicles that can move faster than people, machines that can hurl projectiles more powerfully and accurately than people.
Yet, we still love to watch human athletes do these things BECAUSE WE, TOO, ARE HUMANS. Art, sports, dance, music, cooking are humans doing things that are defined by our shared human limitations and individual, lived experiences. THAT’S why we connect with them so deeply and readily. So some things must, and I believe will, be reserved for humans, or they will have no meaning at all.