Censorship is still censorship, and in the case of LLMs they basically lobotomize the bots. Ai filters are a horrible idea unless the target audience is children.