I mean, you can argue that if you ask the LLM alerting multiple times and it gives that answer the majority of those times, it is being trained to make that association.

But a lot of these “Wow! The AI wrote this” might just as well be some random thing that came from it just out of chance.