It’s art Jim, but not as we know it.
Generative AI sucks, and Henri Matisse knows why.
Now let me state for the record I am NOT AGAINST generative AI, I have used it countless times for video projects but, this is about the heart of art.
The great man Henri Matisse once said, “Creativity takes courage”. Now, every creative genius is possessed of an extraordinary mind. Great artists have impacted humanity and its culture for centuries, and each of them has their own motivations, reasons, passions and experiences that cumulatively provide the backdrop, engine, method and meaning of their output.
Now art, much like beauty, is in the eye of the beholder. I will be the first to admit I have seen some stunning works of “art” generated by engines but, once you know that they have been generated rather than CREATED they lose their lustre.
Generative AI has made significant strides in the last few years, but when it comes to creating unique pieces of art, it still has severe (for that read total) limitations. Despite the advancements in machine learning and computer vision, generative AI struggles in any way to match the level of creativity that humans possess.
Why?
Put simply, soul. It lacks soul. There’s a reason why we say we “pour our heart and soul” into something. Because the very essence of what makes us who we are, the very life experiences, things seen, feelings, and emotions all coalesce into the output at that specific collection of Planck instants.
Edward Hopper once said
If I could say it in words there would be no reason to paint.
There it is, the REASON we create because sometimes words are NOT enough and so it is with AI. Sometimes binary (or Quanary if we go that far) is just not enough.
Georgia O’Keeffe opined:
I found I could say things with color and shapes that I couldn’t say any other way — things I had no words for.
On originality.
One of the main limitations of generative AI is its lack of originality and whence does originality come but from the HEART, the lived experience. While AI algorithms are able to generate seemingly new works of art, they are limited by their training data and so the outputs are an expression of the limited data resolved down to 1s and 0s. One could argue that we as humans are simply the sum of our experiences and knowledge but that is to miss the fundamental thing that AI will ALWAYS lack, soul and that which is built upon our lives as we grow.
Like the rings of a tree are the intricacies of our soul as we grow.
The algorithms use historical data as the information fed to them to generate images, but this can lead to repetitiveness, as they have seen only a limited range of examples and variations. This results in a lack of diversity in the generated art, making it difficult to create truly unique pieces. Now art is not the sum total of its uniqueness but it is so very much more that the sum of its parts.
Sure, you can ask any engine to generate an image of say “A spaceship lumbering through space” and you’ll get a nice poster-worthy picture
but, ask it to “express the inner angst of unrequited love” and you get this:
Seriously, what the heck? It’s basically parsed the words and used weighting to conclude that unrequited love and the angst it produces looks like the outside of your standard middle-tier two-night stopover motel just off the M25. On reflection, this could of course be argued to be correct but only from an existential perspective and binary doesn’t do existentialism. Nor has it, I suspect, experienced the very setting it has portrayed.
Context, context, context.
So that leads me on to the next challenge, context. In studying theology, I was always taught that context is key to understanding everything else and alas, AI can’t UNDERSTAND the context of the art it creates beyond weighted meaning in the request. AI algorithms generate art based on patterns they have learned from the data they were trained on, but they lack the ability and lived experience to understand the meaning or intention behind the art. So, then in turn they cannot generate MEANING which when all is said and done, is the essence of art. I mean, if you listen to “Edge Of Desire” by John Mayer it means something very deep and emotional (well to me anyway) and the context I apply to the artwork of the lyrics, the music, and the fusion of the two may well be different from what the artist intended (for all I know John might have been serenading his Lasagne before eating it) but, it has meaning and I feel cognitively safe applying my own meaning and worth to it, even if in doing so it is not the same as the artists, it has depth and connection.
As a result of this lack of context AI-generated art can (and usually does) appear meaningless or disconnected from reality, devoid of any real emotion. That is not to say that one cannot look at that picture and think, “I like that house”, or “I think that depiction of an M25 stopover has a certain veritas” but it’s only a generated construct until I know a human has created it, THEN it’s art. A dog doesn’t create art (well, from a certain perspective) in the same way an algorithm doesn’t. So you see, this lack of understanding and context is a significant barrier to creating truly unique pieces of art that are meaningful and impactful.
Drive the process, don’t be process driven
Moreover, generative AI struggles to understand the creative process behind making art. While AI algorithms can generate images, they really do not have the intuition and understanding of the creative process that humans possess resulting in outputs that can feel rigid and formulaic, lacking the spontaneity and individuality that makes art truly unique.
So yes, there has been a lot of progress in generating output that imitates art, but much like life, it still has a long way to go. If I am honest, maybe I am being AI-ist, but once I know that something has been completely created by AI I switch off. Not because it’s lazy but because it disconnects me from the creator. If I DON’T KNOW that it’s generated by an algorithm then maybe I can imbue it with the persona of an imagined artist but even so, it’s not art if it isn’t created by a human and that’s a hill I am willing to die on.
But lie to me, and I’ll buy into it.
The limitations of AI when it comes to understanding context, the creative process, and the need for originality all contribute to the challenges it faces in generating truly unique pieces of art. While generative AI may continue to improve in the future, it is highly unlikely in my opinion that it will ever replace human artists.
Vincent van Gogh said
The emotions are sometimes so strong that I work without knowing it. The strokes come like speech.
There’s no generative AI that can ever say something that profound with context and meaning.
Art is human to human communication at a much deeper level.
What do you think?
