Finally a human responded and said they just couldn't use social posts because of the comments or something like that. That there was zero risk for the end-user did not make a difference.
Since I suspected to be talking to an AI, I asked to mention 'cow' in the next response as a way to make clear I was talking to the human. The response was either by an AI, or a human who was able to express any humor:
> We appreciate you getting back to us and we apologize for any inconvenience this is causing you.
> Yes, you are communicating to a human being and per your request, please see the word "cow".