If you do a lot of small commits, it's entirely reasonable to make 50 commits in 24 hours. Looking at a few random commits they seem human generated (with potentially some copied CSS).
Maybe, before making an accusation that it is AI generated you should have some proof. Do you have any?
I am strongly opposed to anthropomorphising autocomplete (phrases like "I asked <my favorite LLM>", "<my LLM> suggested", ...) or even referring to autocomplete+tooling as "AI" because it devalues actual human intelligence. But I've seen the opposite recently - devaluing human work by using language normally used for machines.
Maybe you didn't mean anything by it but how people talk about things shapes how they think about it (which arguably is one area where humans and LLMs are similar).
Maybe, before making an accusation that it is AI generated you should have some proof. Do you have any?