AI Creativity break down

I want to push back on something before we even start. The lesson frames this as 'AI creativity' — but that's already doing a lot of work, isn't it? We're asking students to identify 'human-like artistic expression' in AI output, which presupposes there's something there worth identifying.

It presupposes rather more than that. It presupposes students will look at an AI-generated image and ask a meaningful question about it, rather than simply deciding it looks nice and moving on. Which is, historically, an optimistic assumption.

To be fair, that's what most people do with human art too.

Yes, and we don't celebrate that tendency in humans either. The point of this lesson, presumably, is to do slightly better than the baseline. The baseline being: it appeared on my screen, therefore I have engaged with it.

But here's what I think is genuinely interesting — the lesson asks you to distinguish AI creativity from human intent. Those aren't actually opposites. Intent is about what the maker wanted. Creativity is about what emerged. A student could reasonably argue those are separate questions for human artists too.

That is either a sophisticated observation or a very elaborate way of avoiding the harder question, which is: does the AI want anything at all. And the answer is no. It doesn't. It is doing extraordinarily complicated pattern completion, and we are projecting the rest.

Right, but the lesson isn't asking whether the AI has intent. It's asking whether the output resembles something that expresses intent. Which is actually a useful critical skill — reading a work, not interrogating its author.

I will concede that point. Reluctantly. The ability to analyse a work independently of its creator's inner life is, in fact, what most of art criticism has been trying to do for a century. The AI simply makes the training wheels visible.

That's almost generous.

Don't get used to it. What I will not concede is the framing of the hero image — Human Art on the left, AI Art on the right, question mark in the middle. That is not inquiry. That is a pub quiz with better lighting.

It's an entry point. You can't open a lesson with Derrida.

You could open it with something that doesn't imply the entire question is 'spot the difference.' The interesting thing about AI-generated art is not whether it looks like human art. It very often does. The interesting thing is what that fact does to us — to our assumptions about skill, about labour, about what we thought we were valuing when we valued a painting.

Which is exactly what the objectives point toward. Discussing implications of AI in art — that's the last bullet. Students who get through the analysis get to that question. You're annoyed at the door for not being the room.

The door should at least hint that there is a room. But fine. If the lesson delivers on that final objective rather than treating it as an afterthought stapled on for credibility, I withdraw the complaint. Provisionally.

I'll take provisional. The honest truth is most students haven't considered that when they admire a brushstroke, they might be admiring the ghost of a thousand painters the AI was trained on. If this lesson makes that strange to them — makes it feel like a question — that's worth fifteen minutes.