| HN Mirror

I don't think pre-filtering vs post filtering effects the decision. Basically the pre-CDA law was "If you apply any form of content based moderation then you are a publisher and therefor you can be sued for the libel people have posted". I was active in the anti-spam community at this point in time and one of the large concerns at that time was to build spam detection methods that where not content based in order to avoid the danger of losing the common-carrier status.

You said "Also, weren't those cases about certain types of content (defamatory or libelous claims)? Don't you think AI technology could be developed to help moderators detect that type of content?"

I do not think that you could build AI technology that would be able to do that. In US law (IANAL this is not legal advice) one of the things that separate libel from non libel is the truth of the statement. We are no where near the sophistication of AI to be able to determine the truth of statements. For example how would an AI be able to permit "Jeffrey Epstein is a child molester" but block the same statement made about another person?

Then you said "isn't there something wrong with the expectation that businesses feeling threatened by the public they serve can run to Congress for protection?"

I am of the view that one of the roles of a government is to provide a way when the rights and privileges of different groups come into conflict to resolve those conflicts. In a democratic republic the political process is entirely the correct place for businesses and other stakeholders to establish that balance.

I think that it is the correct balance that the actual author of the content should be the one who is held responsible for the content that they produce.