[Majromax]

But the NYT can carry liability for letters to the editor published in its dead tree format -- see https://www.rcfp.org/supreme-court-will-not-hear-letter-edit... as an example of a local newspaper being held liable for letter-to-the-editor-published defamation.

The CDA draws a bright line between content "authored" by a firm and content "made available." In practice, that line is fuzzy.

As a hypothetical example, Twitter probably should face liability if it took a random tweet (say) accusing Bezos of pedophilia and made an editorial decision to promote that tweet to all its users, but it could still plausibly claim that it was just making the content available.

It's a complicated topic, and I don't know where the best balance lies.

[xxpor]

The tweet promotion is an interesting point, but the letter to the editor is easier IMO. It's assumed that a human has read and selected the letter to the editor, which is why they'd have liability. For the promoted tweet, my first reaction would be to say, if a human affirmatively promoted it, they'd be liable. If it's pure algorithm, they wouldn't be if they took it down when served a notice.

[simonh]

That’s not the current situation under Section 230. You can even re-tweet or forward content posted by someone else and not be liable. Only the original author is liable. This is sensible baca use otherwise all sorts of innocuous relaying, trending and categorisation activity normal of forums and social media that affect the scope and visibility of posts could trigger liability.