I found it curious that chatGPT security protocols prevented me, in a japanese language learning exercise roleplay, from telling a fictional conbini (711) emplo

I found it curious that chatGPT security protocols prevented me, in a japanese language learning exercise roleplay, from telling a fictional conbini (711) employee that they have nice hair. ChatGPT said it was potentially offensive and would not continue the roleplay. I primed chatGPT for the exact same scenario, using the same words, but replaced japan for cuba, and chatGPT had no problem letting me compliment the conbini store employee's hair. So truly it does not have a unified view of human ethics. It treats each culture within its own context. Even for its own self-moderation.