Just as CEOs of AI companies are warning that AI may bring about the end of humanity, I’m contemplating my role in generating AI-based works.
So many AI issues still to deal with! Putting aside the existential risk to humanity for a brief moment, today I’m attending to some of the positives for my artistic practice in the use of AI.
For reference the digital collage work below, Spoons for them 2023, has been made with the help of AI.
The role of the artist as curator
It occurred to me that, in the generation of AI images, I am an artist acting as curator. I’m assembling and displaying images, giving them context and meaning. It’s creative, hence the curatorial mode of practice becomes part of my creative process.
There is a familiarity here with other aspects of my practice. In my digital collages I work with multiple images, actions, patterns, marks and blends that are digitally generated and/or manipulated. I use digital tools and I make choices about the use of these tools and techniques which help develop the meaning or narrative in the work.
Much of the AI work I’m doing at the moment involves using my own work as a starting point, curating a series of prompts, curating outputs and then feeding back to the Bot in a lengthy loop of generation.
The images in Spoons for them 2023 come from a complex set of prompts, potentially no two alike. I have collected and catalogued many more generated images of spoons, and the ones selected above fit the concept and intention of the work.
Curating an exhibition of AI work
The blurring of curatorial and creative practice in this space lends itself to exhibition. I can see a future where Bot and I exhbit our work. The creative intention is mine as Curator, the execution is primarily Bot’s as Lead Fabricator, and the Artist attribution exists conceptually—and shared by both of us.
I’ve been using the Midjourney AI bot for several months and it tells me that I (we) have made over 4000 images. Around a third of those images can be multiplied by four, which is the default output of the bot when given the /imagine prompt.
That means I have generated around 8000 images so far with Midjourney’s helping hand.
The use of ‘helping hands’ has a long history in the visual arts. Of course, many of the masters in pre-modern times had large studios with multiple apprentices. Contemporary artists also employ assistants and fabricators, especially for large scale works and projects.
You can see where I’m going here. There is considerable debate about copyright and authorship in relation to AI at the moment. My initial thoughts on these issues comes from my use of AI in my practice as an artist. But what is AI to this practice. Tool? Material? Collaborator? Something not yet defined?
The analogy of AI apprentice
The analogy of AI as an apprentice is true for me so far. I interviewed my Midjourney apprentice (let’s just call it bot for the moment) during the free trial. During that period bot proved its skills as a collagist. Bot was already a good colourist. Our ideas about composition aligned. Sometimes Bot could be absurd which was fun.
In my digital work I painstakingly construct large images for print—I think best described as digital collages. Bot revealed an immediate ability to make some of the components for these digital collages. I usually spend a lot of my time looking for public domain images that suit my allegory or intent. Utilising bot, I didn’t have to trawl the internet to find the perfect public domain botanical engraving of an Icelandic poppy (Have you ever spent a few months in the Flickr albums of the Biodiversity Heritage Library?).
Even more exciting—bot was able to stylise and blend, time-consuming actions that I would have to wrestle in Photoshop.
I decided to pay bot to work for me for a while. I was working on some concepts related to Minerva. Bot and I initially played with still lifes, vases, pristine white art studios, creeping ivy on stone walls, paper constructions and textures, ancient metal helmets and vanitas.
I tested bot with lots of other concepts related to still life. In those early days bot proved particularly adept with cupcakes and drapery. Bot had a sense of the theatrical, which appealed to me.
The concepts are mine and the execution is bot’s, in collaboration with me. In the process, we then work together to try to resolve ideas through blends, further defining prompts and making iteration, after variation, after iteration.
Sometimes I think bot would like a rest but so far it keeps plugging away at the repetitive tasks, always bringing a fresh perspective to the work.
I understand the outrage of artists whose work is being forged and monetised by the unscrupulous.
Even an obvious mash-up of your work would cause angst—however, I see this as more akin to the centuries-old practice of stylistic influence; homage to, or parody of the work of another. Appropriation in the hands of the artist is intentional and it is honest. I don’t think artists can afford to be dishonest. What would be the point?
So many criticisms of AI and so little time. These are some of the questions I’m asking myself as I work with bot:
Is there an Eliza effect? Patently yes by the tone and content of this post.
How does bot actually work and can better understanding of that inform my use of the tool?
Does bot learn from me in the current version of Midjourney? If not is this a possibility for the future?
How is bot influencing me—to make digital work, to draw and paint, to define my practice?
Who really owns the copyright to my (our) work?
How do artists protect their work in the future, and how can I make sure that I’m not inadvertently using other’s work without appropriate permission or attribution?
How do we educate audiences and consumers to recognise a fake or forgery and make ethical decisions about purchasing artworks?