A site like Twitter or Facebook that has millions of users contributing content can edit the feed to convey a particular message by only showing posts that fit. Just like an author who chooses particular quotes that fit an article, but on a larger scale, and the "article" is now made only of quotes.
IOW, the message can be crafted by the platform even if the words are provided by the users.
There is no article trying to reduce a many to many conversation to a simpler construct in order to justify regulating it is poor analysis.
Platforms may or may not be slanted people can pick a different platform if they don't like their current ones slant or make their own. People ought to have no right to have a platform to be neutral in what communication they facilitate between their users and any attempt to enforce that is inevitably going to devolve into being slanted towards those with the juice to hire the lawyers to suppress speech.
Whether you call it an article or something else, the feed as a whole, if chosen in a biased way, represents the point of view of the selector.
Just as an article made of quotes would.
And if the message in the feed/article is libelous, it's shameful (though perhaps legal) to hide beyond the argument "I was only quoting other people."
IOW, the message can be crafted by the platform even if the words are provided by the users.