Despite many billions of dollars collectively thrown at the problem, "AI" remains stubbornly uncreative. A version of Google's Gemini did score at a gold medal level for the 2025 math olympiad, though its significance is contested, naturally. In any case, it's clear that the conventional approach isn't enough to crack the code on generating something interesting or new. This is particularly the case in art. It certainly makes a strong case for human creativity existing at some magnitude higher than a conglomeration of data, though I wouldn't say that was an immediately evident conclusion at all.

I had some vague thought that AI don't have individualistic views and experiences to the extent that humans do, and maybe that would give its output more personality.. but there is a chasmic difference in the output from a competent artist and AI given a more specific and "interesting" prompt anyhow, so presumably that wouldn't be enough.