If the goal is to prevent the ambiguous harm of "editorial effect", then I don't think it is fair to remove all platform immunity and punish the platform every

If the goal is to prevent the ambiguous harm of "editorial effect", then I don't think it is fair to remove all platform immunity and punish the platform every time it fails to implement the correct, government-approved "editorial effect" instead of its own. It seems like the proponents of such a rule change are being underhanded, thinking "We can't ban companies from having a political bias, so we'll say that if the company has a political bias (i.e. any editorial/content policy), it becomes liable for any libel, or scams, or threats (written in any language) that appear anywhere on its platform". I might support a narrow form of this, though, which says that if a platform doesn't let you opt out of (legal content) filtering/re-ordering of content, then the platform has profited from you receiving messages with an unwanted bias (i.e. commercial speech), and therefore owes the user a small amount of statutory damages each time the user suffers some harm.