Picture this: you make art for a living. You pour your blood, sweat and tears (hence the term ‘the struggling artist’) into every commissioned visual. One day, you find out about a new competitor on the market. This is nothing new, only this time round it is a computer programme that can do whatever you’re doing within seconds and at no cost!
In face of a seemingly undefeatable competitor, namely A.I. generated art, the brewing bitterness and anxiety among creative communities is completely understandable. It is easy to deduce what companies would turn to in such a profit driven world.
A.I. generated art is not a recent development. However, the newest models such as DALL-E2, Midjounrey and Stable Diffusion feature functions that turn simple written prompts into complex looking artworks. Generally speaking, A.I. generators are trained on databases with billions of images and existing artworks, meaning a lot of living artists are unknowingly contributing to the development of A.I. art generators. A.I. passively absorbs data in the images such as their subject, composition and style, then collages them together to produce a new piece. The implication is that anyone, regardless of skill or talent could produce art that would otherwise take hours, if not days to complete, not to mention artistic skills and sensibilties that take a lifetime to foster.
Some raise the argument that AI generators work no differently to how humans are influenced by other artists and artworks. Indeed, Paul Gauguin’s painting Spirit of the Dead Watching took inspiration from Edouard Manet’s Olympia, while Van Gogh’s art was hugely influenced by his post-impressionist contemporaries. What the aforementioned argument fail to account for is the creative process that living artists go through. Most artists’ journeys start by imitating others. They do, however, develop their personal style and artistic concepts as time progresses. A.I. generators, on the other hand, only exercise what can at best be referred to as ‘combinational creativity’ by formulaically mashing together images that a few keywords signify – there is no real creativity involved.
A.I. generated images and the mechanisms behind it completely undermine the humanity in art, subsequently devaluing the essence of art. Interpretations of what constitutes art vary. But to a lot of living artists, art is defined by its creative process as much as the end product, which is a view that I too personally hold. Author and illustrator Rob Biddulph makes the significant point that a human artist adds emotion and nuance into their art. Art could be deeply personal. Complex culminations of personal experiences, values, visions etc. shape human creative production. This exposes the inhumanity of A.I. generated art as computed systems fail to grasp or express these human concepts. Such observations quickly usher in criticisms that A.I generated art is anti-artist and fundamentally the antithesis of art.
There’s also the question of ethics. Some rush to the defence of A.I. systems such as DALL-E2, expressing that it is merely another tool to further inspire creatives. While these tools hold potential to inspire, it also invites abuse. Instances of abusing these A.I. generators include the production of photorealistic child pornograhy or other non-concensual NSFW images. Furthermore, A.I generators don’t credit their ‘inspirations’. When it comes to A.I. generated art that carries an uncanny resemblance to an identifiable, copyrighted original source, concerns over intellectual property infringement arise.
Distressed artists are increasingly voicing out fear that A.I. would put them out of work. There is some solace in the fact that we as humans are more inclined towards the personal and authentic. At present, A.I. generators cannot accurately replicate artworks or replace the personal touch of art. However, as A.I. continues to improve, the line between human art and A.I. art will start to blur, if it hasn’t already.
Image Credit: Kelly Sikkema on Unsplash