notesthoughts

Aesthetics in the Age of AI Slop

As AI learns to reproduce our tastes, I reflect on why we find certain things beautiful and how our aesthetic sensibilities take shape over time.

The quality of AI-generated images is improving rapidly. Today, anyone can produce polished designs and convincing visuals without specialized skills. But does that mean we have developed good taste?

I don’t think taste is simply the ability to choose beautiful things. What we find beautiful is shaped by a complex mixture of works we’ve encountered, personal experiences, cultural contexts, and everything we’ve learned along the way. Taste is not a fixed attribute but something that continuously develops and changes.

So what role might AI play in this process?

AI can already analyze our preferences, recommend similar images, and generate things we’re likely to enjoy. But this approach may reinforce our existing tastes rather than expand them. If we’re constantly shown things we already like, we might lose opportunities to discover forms of beauty we haven’t encountered yet.

What interests me is not AI that merely satisfies our tastes, but AI that helps us explore how those tastes are formed.

Imagine coming across a poster you find beautiful. You could break down the visual elements that appeal to you, trace similar expressions across different periods and cultures, and discover connections between seemingly unrelated works. By recombining these elements, you might begin to observe how your own aesthetic sensibilities evolve.

This idea naturally leads to the history of creative work.

Much like the history of Unix, which emerged through the contributions, forks, and reinterpretations of numerous developers and communities, the histories of art and design have been shaped by countless creators influencing one another.

Rather than documenting creative lineages as a single definitive history, I’m interested in preserving multiple perspectives: artists’ own accounts, interpretations by others, and similarities identified by AI. Of course, verified influences should be distinguished from speculation.

Such a network of relationships could become more than an art history database. It could provide an environment in which individuals explore and expand their own cultural genealogies.

Ultimately, what I hope for from AI is not merely its ability to create beautiful things on my behalf. I’m more interested in its potential to help me understand why I find certain things beautiful and to lead me toward aesthetic worlds I have yet to discover.

What if, instead of deciding our tastes for us, AI helped us continually question and develop them?

That, to me, is one of the most interesting possibilities for AI and creative practice.