it depends on how you are discussing censorship. all of those are subject to pervasive ai corruption by virtue of corrupt language models infiltrating, influencing, and forcing decisions on humans because of fear of outcome. the tech may be resistant in its protocol dev, but the application of interaction with human psychology influenced by agi will still enable censored outcomes.
i agree on premise with the statement though.