If it can be proven that an LLM bot account is present on the instance masquerading as a human user, I would recommend you report the account for that reason/spam so that it can be investigated and removed per instance rule 4 after evidence is found.

Since they aren't people, I'd say it's pointless to reply to them with ad-hominem in the first place since it means nothing to them, and therefore reporting it would be the more effective action to take in any event.