My new year’s resolution is to stop posting about AI. Here’s my last one. In 2023 I started speaking publicly about the influence of AI on music. It was part of my job to figure out how to respond. Anyway, I noticed that there was this building movement towards criticizing the ethics of it, primarily centered on the content used to build the databases they were built out of. Essentially every ethical question about generative AI centered on how artists were going to be compensated for their work. I have said, on numerous occasions that I didn’t consider this to be the most important issue. It certainly matters.

However, none of this discussion ever touched on what I felt was the bigger ethical problem. Perhaps because it is an aesthetic problem. And perhaps, more importantly, generative AI was seen as transforming aesthetic problems into technical problems. Regardless, for me, the issue was always the fact that they were engines that enabled the rampant growth of ugliness. They are ugly to me and the internet and everything else that they touch in our world (ads, posters, event invitations) is now uglier.

Some will counter that ugliness is subjective. So are ethics in an uncaring universe. We all know it is ugly. This is why everyone calls it slop. Still, perhaps more importantly, I think generative AI has given me a new definition of ugliness that connects aesthetics to ethics in a way I hadn’t considered before. If humans have one virtue, it is that we are capable of creating new information. Not, variations on a theme, but genuinely new information that is truly anti-entropic. Generative AI gives us an objective definition of anti-art. You cannot train it on its output without poisoning it. Human art has value we have not previously understood. Real art–whatever it actually is–is a fundamental ethical good that expands and complicates our experience of the world. AI slop simply cannot do this.

I wrote about this from a different perspective in my dissertation–arguing that nostalgia was only damaging to a culture when it produced the kind of feedback loop that only reproduces itself. If it becomes a grotesque parody, or an amalgamation of experiences, or what the creators imagined a time period to be, it can become the basis of something profoundly new.

Generative AI systems are based on the false assumption that an interpolation of extant art is the only basis for new art. It isn’t. What they enable is a flooding of the world with low quality slop. We sense it. It makes us feel sick to look at it. How many of you can’t explain why, but feel gross when you see AI content? This feeling that something is wrong… this sense of horror at the uncanny (from the German unheimlich, literally the un-home-like, as opposed to the “homesickness” of nostalgia)–is perhaps a kind of preservation instinct we have against the pollution of thought.

Our culture hasn’t historically treated artists with the value and respect of other professions. It has, personally, felt difficult to justify what I do in a larger cultural context. I do wonder if the damage done from generative AI, when we begin to quantify it, will finally provide that justification to the broader political economy.