where it is insensitive to engage in a topic about one gender or class of people, but will freely joke about or denigrate another by simply changing the adjective and noun of the class of people in the prompt
the US left leaning bias is around historically marginalized people being off limits, while its a free for all on majority. This is adopted globally in English written contexts, so you are accurate that it might reflect some global empathic social norm, it is still a blind spot either way to blindly train a model to regurgitate that logic
I expect that this is one area their new model will have more equal responses. Whether it equally shies away from engaging, or equally is unfiltered and candid