I agree. The problem with censorship is that then it needs to be finetuned to different cultural contexts. Otherwise it ends up imposing American standards on everyone. We already see enough of that, like TV having to adopt american diversity quotas, even when those quotas make no sense in places with different demographics. This is relevant since AI censorship has adopted the particular american sensitivity around ethnicity, and applies it in countries where people aren’t as obsessed about skin color.
It would be simpler to just cut out the thought policing, but it seems the developers are determined to act as gatekeepers for what information the plebs get to see. So here we go: customized “those words shall not be spoken” filters for each culture.