The main thrust of the Executive Order is its attack on 47 USC § 230, the law that underlies the structure of our modern Internet and allows online services to host diverse forums for users’ speech.
Let’s break down section 230 a bit to understand both the EO’s intended impact and erroneous legal footing. Section 230 grants online intermediaries broad immunity from liability arising from publishing another’s speech. It contains two separate and independent protections.
Subsection (c)(1) shields from liability all traditional publication decisions related to content created by others, including editing, and decisions to publish or not publish. It protects an online service for hosting user-generated content that others claim is unlawful.
Subsection (c)(2) is an additional and independent protection from legal challenges brought by users when platforms decide to edit or to not publish material they deem to be obscene or otherwise objectionable. Unlike (c)(1), (c)(2) requires that the decision be in “good faith.”
The protections in (c)(1) and (c)(2) are independent. But the EO reads them together, calling (c)(2) a qualification of subsection (c)(1) thus requiring that all publication decisions be in “good faith.” And the EO tasks government agencies with defining “good faith.”
This means that for a single act of editing user content that the government doesn’t like, a platform would lose both protections under 230 and could be subject to onerous private litigation for thousands of completely unrelated publication decisions.
So the EO purports to give the Executive Branch powerful leverage to force platforms to publish what it wants them to publish. But even if section 230 permitted this, and it doesn’t, the First Amendment bars such intrusions on editorial and curatorial freedom.
The Supreme Court has consistently upheld the right of publishers to make these types of editorial decisions. While the EO faults social media platforms for not being purely passive conduits of user speech, the Court derived the First Amendment right from that very feature.
In its 1974 decision in Miami Herald Co v. Tornillo, the Court explained that 
Courts have consistently applied this rule to social media platforms, including the 9th Circuit’s recent decision in Prager U v. Google. Just last term in MCA v. Halleck, the Supreme Court rejected the argument that hosting the speech of others negated these editorial freedoms.
@EFF filed amicus briefs in each case, urging that very result. eff.org/document/prage…
eff.org/files/2018/12/…
eff.org/files/2018/12/…
But even if the First Amendment were not implicated, the President cannot use an Executive Order to rewrite an Act of Congress. In passing 230, Congress did not grant the Executive the ability to make rules for how the law should be interpreted or implemented.
The EO appears to have been prompted by curatorial decisions the President personally dislikes, and not a broader concern for freedom of speech and the press. We should all be concerned that it will be implemented in an equally retaliatory manner.
Despite the lofty praise for “free and open debate on the Internet” in the EO, this Administration has shown little regard, and much contempt, for freedom of speech and the press. I’m skeptical the EO will actually advance the ideals of freedom of speech or be justly implemented.
The final version of the EO is still awful but a bit different than the draft broken down in this thread. The final version does not itself mush (c)(1) and (c)(2) together, abandoning the draft's "qualifies" language. Instead, the FCC is to opine on the connection between them.
So the EO doesn't specify what effect on c1 immunity an adverse good faith finding will have. But it wants the FCC to figure that out. There are some other differences as well. But nothing else I think changes anything in my thread. The final version here whitehouse.gov/presidential-a…
