The AI art ouroboros
Why artists are worried
“If we, as humanity, rely solely on AI-generated works to provide us with the media we consume, the words we read, the art we see, we would be heading towards an ouroboros where nothing new is truly created, a stale perpetuation of the past.”
That quote comes from AI Art and its Impact on Artists, a new paper published by a group of artists and researchers and picking up traction on social media over the past week. The entire thing is worth a read if you have the time. It delves into the copyright questions we’ve previously examined and outlines the many reasons artists, particularly visual artists, are anxious about AI developments.
The paper was written by, among others, contributors from the Distributed AI Research Institute which is “rooted in the belief that AI is not inevitable, its harms are preventable, and when its production and deployment include diverse perspectives and deliberate processes it can be beneficial”. That includes founder Timnit Gebru who was controversially fired from Google’s AI ethics team last year.
The ouroboros analogy, i.e. a serpent eating its own tail, and versions of that concern come up a few times in the paper. It’s a neat visual of a future dystopia where no more human input is required and audiences just keep getting different versions of the same thing (hey Marvel and Star Wars 👋).
It also addresses a pet peeve for anyone trying to advocate for human input when discussing generative AI: the anthropomorphization of AI tools. “Ascribing agency to image generators diminishes the complexity of human creativity, robs artists of credit (and in many cases compensation), and transfers accountability from the organizations creating image generators, and the practices of these organizations which should be scrutinized, to the image generators themselves”. However, it’s worth noting that many AI art community evangelists tend to avoid anthropomorphization when discussing tools, instead lionizing their own achievements in using said tools.
The paper explores biases in gen AI, citing a great feature on Senegalese artist Linda Dounia, who says, “any Black person using AI today can confidently attest that it doesn’t actually know them, that its conceptualization of their reality is a fragmentary, perhaps even violent, picture”.
The concept of ‘data laundering’ is also addressed, the theory that companies can stay on the right side of copyright law by getting datasets obtained by nonprofits for research. That this degree of separation strengthens their fair use argument. Add to that the thin, sometimes nonexistent, ‘membrane’ between academia and the tech industry, and a pretty bleak picture is painted.
The whole paper is a good base text for anyone trying to argue for the value of human creativity. Practical solutions to these problems revolve around stronger regulation, greater transparency and accountability around data collection, and stronger union representation for artists. All fun, and very sexy, things that will no doubt be addressed as quickly as they need to be.
Small bits #1: Weird AI please, heavy on the weird
It’s been a while since we had a creepy AI-generated video nightmare in the newsletter so here’s the Friends opening “as created by AI”, with perhaps more alien content than you would expect. Made using “Midjourney and then Pika Labs” according to the uploader. The current sweet spot for virality (it did numbers on Reddit) remains showcasing the massive flaws of text-to-video AI, probably this creator’s intention.
Small bits #2 ChatGPT and AITA
There’s a current mini-trend for making ChatGPT create posts for the Am I The Asshole? Subreddit. This version did well on the ChatGPT subreddit and this guy wrote a blog about going viral with his faked AITA post. The format, someone cluelessly describes being a complete asshole/being the victim of asshole behavior and gains virality through incredulous responses, is well suited to LLM stunt posting. Any Quite/Very Online person has read an AITA in shock, briefly thought it may be a complete fabrication, and then shared it anyway because it’s too good not to. Off Reddit, AITA screenshots are regurgitated for numbers all the time. The shaggy-dog-stories-on-social industry is definitely open to destruction/boosting by generative AI.
Small bits #3: New AI discourse concept just dropped
This was a throwaway remark in a discussion around some marketing speak about a new ‘ecological’ AI agency, but it feels like a good descriptor of a lot of AI-related projects at the moment. Where AI ‘speaks’ to grab our attention, but humans are very much doing the work of a puppet master.