3D rendered image of a telescope on top of a building, looking at flying spheres of streetlamps. The roof top itself is styled after a circular labyrinth.

Many are willing to die for their art, but few are willing to learn to draw?

AI – TL;DR: I’m very very sceptical – and for the purposes of this blog, i’m talking about LLM/Generative AI in its use-case for creation of content.

We can possibly imagine a future implementation of AI based on ethical underpinnings, but what we have right now isn’t it – the power and processor usage is absolutely atrocious, their approach to data protection and content ownership is abysmal, and it is fully based on laundered content with no consent or opt out – especially in the world of visual arts and illustration.

Artists and Academics have always remixed prior works?

I don’t buy the idea that AI usage is the same as an artist working with found materials, sampling in music or an academic re-synthesising prior work because those things are curated with intent and understanding of those sources, and ultimately, a human name sits next to the work who is both responsible and accountable for it. AI obliterates accountability and replaces it with a veneer of bland objectivity.

So while I do think Roy Lichtenstein is a git who made bad copies of Russ Heath’s art with the justification that elevating comics from low to high art was transformative, he at least got the brushes out and made conscious decisions about what to copy and how – and can be held accountable for attempts to justify that and the influences traced back. That’s a cultural conversation about the nature of art and creativity that can be had, but with AI, everything that is interesting about that conversation is removed: It’s stolen – who from? – don’t know – why? – don’t care.

It’s just another tool shaking up art, like Photography or Digital Painting?

I agree that, as with the advent of Photography and Digital Painting, AI will result in new developments and the evolution of art as a reaction to it – and yes, reactions to Photography and Digital Painting were much the same as to AI now, especially when first introduced.

But we are not the same: Photography requires knowledge of composition and framing, lighting, timing and creates a unique image based on where you are standing and the decisive moment you choose to press the button. You own the tools, and the result is unique to your location in space and time.

Photography is a medium for expression. AI is a means of production. More on Marx coming right up.

“The underlying purpose of AI is to allow wealth to access skill
while removing from the skilled the ability to access wealth.”

The image accompanying this post is most certainly not AI – it is made by me, and is a combination of photography, rendered 3D and digital painting: I’m not anti-tech and have no issue with procedural generation for example.

It’s inevitable – if we don’t use it we will be outcompeted by those who do?

The argument of inevitability is one used by folks in a cult – and it’s being pushed by exactly the same people who were pushing NFT’s and Crypto-currency. I don’t doubt that people not using AI will be at a disadvantage to those who do – I’m disputing its ethics not its utility*.

But if it is an inevitable tidal wave of technological innovation, that fact does not oblige us to accept, and by resisting, we can perhaps influence the direction of travel even if we can’t prevent it.

The word Luddite is used to imply a small minded resistance of progress, but keep in mind that the Luddites had a number of very solid points and I don’t think they were ultimately proven wrong. They actually had no objection to machines or technology and were frequently very capable of using them, but they were concerned with technology being used in a fraudulent manner to undermine good labour practices or the quality of the goods produced.

So it goes.

The content is on the internet and that’s fair game?

Search engines have been trawling content for decades now, and in most cases this is welcomed by website owners who understood that the business model was basically, if they have good content, Google and Bing send traffic to them, and along the way both parties make some money from advertising.

That assumed contract has now been torn up – see this post from Andy Oliver of Broken Frontier. Their writing has now been laundered and presented above the search results (and presumably also in apps who knows where else) lowering the likelihood of traffic to their site. The enshittification process enters its next phase.

But I wonder what Google and Bing think the bottom of the barrel looks like here? Those producing good art and writing will find ways to protect it, and all that will remain to crawl will be grey goo generated from the same models to be fed right back in.

Suffering for your art?

I’ve seen the argument that artists should make art for its own sake, and at the same time belittling the value of art, writing and content creation generally, as a justification for refusing to pay and as a justification for the use of AI, which is an attitude beneath contempt.

Whatever Zuckerberg’s opinion, people who create content should be able to dictate the compensation they receive for your viewing of it, even if that compensation is as limited as having you visit their site (hi! welcome) and the joy of number go up in their analytics.

The quote at the top of this post originates I believe with Simon Munnery (but this and variations on the theme have also been attributed to others, including Banksy). Interestingly when I asked AI savvy people about the quote, one of my correspondents AI had made some interesting attempts to explain the meaning, but zero attempt to identify the author.

—————————————-

  • Actually I probably do want to challenge the utility somewhat. When I mentioned that AI writes with a veneer of bland objectivity, it also does so with an assumed authority even when it does not have a clue. That is actually quite dangerous.

    I challenged friends to decipher a coded message, and one believed that they had solved it using Chat GPT – which had both suggested the plaintext but also the means of encryption. It was entirely (hilariously and obviously) wrong on both counts, but as I say, was wrong in an objective and authoritative tone.

    More on my Cipher Disk project later, but if you want to try some cipher cracking, here is the text:
  • BCFTE UMLQR JSYS NOR ZN
  • GBKDO FOILX IBB EEX FOSP
  • UUKQB JTBTW QKA OMCKXN SIL HRXSYC FEJ
  • MQOMF CVXFS TNXX HUG FGZ
  • GZGKB SPLMV PYZ EFJQCIZ
  • CGVCD AJMDH RNPO W PUB VLN XDAD ZFD

Update 2025:

2025 needs to be the year we retire ‘lmgtfy’ or ‘google is your friend’ as a response to people in niche interest groups when new folks ask very frequently asked questions.

A decently curated welcome page could solve that, and Google hasn’t been anyone’s friend, certainly not yours, for years now, and what people want when they arrive in such groups is a connection with humans who share their interests.

Meanwhile:

* https://aramzs.xyz/thoughts/dont-post-ai-at-me/
* https://www.digitalpublic.io/ai-invasive-species/