GenAI is theft.

That used to be the start of this post. Then this happened.

What a world…

This reporting fills me with abject horror and dread for the world we live in.

Where those without values of their own care nothing for the values of others, nor for the value of others.

So, it is more than theft. It is theft and destruction…sometimes of rare and irreplaceable books. Even when pulled from PDFs, online repositories, etc… for any form of writing, be it creative or otherwise, some argue fair use, but copyright law and the good ol’ Berne convention are pretty clear. It doesn’t matter where you put it, it’s still copyrighted. Anthropic even lost a lawsuit. Because they are thieves. They thieved. They done thefted. It was for $1.5 billion which, at their current revenue of around $82M a day, is just a little over 18 days of revenue.

18 days.

Art, writing, music…

There is nothing to generate from without stealing. Steal - and, being directed to steal - is what these tool systems do and have been instructed to do.

This is not in dispute, but in case search engines feel challenging:

Environmentally, GenAI is disastrous.

And, of course, jobs…

Artistically, GenAi is devoid. Just as ‘content’ isn’t writing, even though writing can be content.

When I started this post, this also hadn’t happened yet:

What a world.

Our richest man is suing a state because they don’t want his AI to make non-consensual deepfake porn. Oh, and of course child porn…its specialty, I guess? In fairness, apparently, xAI is suing a user over this…one user.

I am opposed to the use of GenAI in the arts, full stop.

AI ads gross me out. AI ‘graphic design’ is particularly shit and has apparently taken over the ‘flyer’ space.

I don’t want to patronize it. It saddens me to see the shortcut to a thing. Creativity is not an outcome, it is a process. A craft. I do feel disdain for its use.

Apologists will hem and haw its place. Are there use cases where these toolsets and LLMs are valuable?

Of course.

Research? Potentially invaluable, already. (But, interestingly, is there any compensation for those that published the research that trained the AI that makes billions for others and nothing for those with the talent to have done the research? It’s a fair question, although you can also go back 20 years when Google and, then, Facebook, murdered journalism (like the vampires of old, they were invited in…) to see similar patterns. It’s a playbook.)

Security? Indubitably. Agentic models can detect hackers and hacker agents faster than any human.

Medicine? Apparently, yes. Especially in early cancer detection.

But art?

I won’t argue the aesthetics of the things…I think of the hellish visions of many horror GenAI accounts and yeah, they can look amazing. But the “I created X in what would take a real artist Y time” is a fallacy. It’s like saying, I stole this bread faster than a baker could bake it.

The writing is still relatively shit, but that will also advance over time. I look at advertising copy and as it has mostly descended to crap, that’s what AI has learned on, but these models are stealing from the greatest writers in history. They have been directed and supplied by people who think that’s okay. Maybe you think that’s okay. I don’t think that’s okay.

Having something isn’t making something. I have experimented with GenAI. I work in a world where AI tools are very useful. I think of Canva using AI on an uploaded ad or video to resize it for 23 different formats. That’s amazing for production time. It makes a lot of sense for what Dave Trott would refer to as the applied arts. But, for the ‘pure arts’? For music? For writing? Dance? Theatre?

That’s a no for me.

Art, as Ethan Hawke says, in the following lovely excerpt, is sustenance.