One of the dangers of completely relying on AI is that someone else gets to decide what we can generate with their models.

Just ask Microsoft about Tay. On the one hand, I understand why you want some censoring in your model, on the other, I think it also cripples your models in unexpected ways, I wonder if anyone's done such research, compare two models by the same source training data, one with censoring of offensive things, the other without. Which one provides more accurate answers?