When Google for instance penalizes a site for being slow; or for being spammy; or being identified by the algorithm as likely to be disliked by most users, and not be what they are looking for; or for having spyware; or for seeming to use "black hat" SEO with irrelevant keywords...
...are these actions "censorship"? What kinds of shaping of google results are censorship and what kinds aren't?
The results are inherently shaped, they only exist because of algorithms that implement choices, there are no "natural" unshaped results.
You have to make a choice of what you prefer but the choice you're making is censoring slander or not.
Which of course it doesn't, a state-of-the-art AI's language comprehension is still extremely rudimentary and would be prone to being gamed by powerful malicious actors even if it were human level.
But let's play along and assume that everybody else's speech is the problem. Citizens don't need MSM fitted, woke muzzles.
The co-opting of internet providers to mete out draconian terms of service that strip the individual of their freedom to express ideas needs to be criminalized.