Former U.S. President Donald Trump is suing Facebook, Twitter and YouTube, claiming he was “censored” by those companies for not extending him the courtesy of consequence-free posting  on their platforms. The complaint, filed Wednesday, alleges the companies used “unconstitutional authority delegated to them by Congress” to deny both Trump and four additional plaintiffs their First Amendment rights. 

A suit that made similar allegations against PolitiFact, Science Feedback, the Poynter Institute, Facebook and Mark Zuckerberg was thrown out last week.

Both seem to stem from a misunderstanding of the First Amendment to the U.S. Constitution’s protection of freedom of speech. The amendment grants citizens freedom from restrictions on speech that come from the government, not private entities. And as anyone who’s taken a media law class can tell you, there’s a whole host of caveats to that freedom that have been shaped by over 250 years of American jurisprudence.

Looking at the complaint, it’s also clear there’s a fundamental misunderstanding of how Facebook makes content moderation decisions. Trump cited the Facebook Oversight Board’s review of his suspension that criticized the seemingly arbitrary nature by which the company removed him from the platform. Another complainant who was temporarily suspended for sharing anti-mask misinformation said her misunderstanding of the rules prevented her from engaging her network on the platform to find her missing brother.

Fact-checkers working with Facebook similarly suffer from a user misunderstanding of how Facebook moderates content. Many Facebook users see any kind of labeling of their content as censorship. Some take it a step further and say fact-checkers are deliberately censoring their content and limiting their speech.

That’s not how the Third-Party Fact-Checking program works. Justine Isola, Facebook’s head of misinformation policy, explained on a United Facts of America panel that fact-checkers use a portal set up by Facebook to attach their fact checks to posts that have been flagged by either Facebook users or Facebook’s artificial intelligence. (Full Disclosure: Fact-checking organizations are required to be signatories to the International Fact-Checking Network’s Code of Principles to be eligible to partner with Facebook).

Facebook uses the work of its fact-checking partners to train its AI how to recognize similar posts independently. Facebook also uses this work to decide, independent of its fact-checking partners, which posts to remove and which posts to limit distribution.

This isn’t a flawless process. Machine learning technology is only as good as its inputs, and there simply aren’t enough fact checks or fact-checkers to help perfect Facebook’s AI.

The resulting animosity over users having their posts labeled or their distribution limited is putting some fact-checkers at risk, with some facing protests outside their offices and others facing increasingly violent threats.

IFCN director Baybars Örsek announced Wednesday the fact-checking community will band together to address the recent spate of harassment against fact-checkers. But the truth is that much of it stems from a misunderstanding of who has the power to do what to whom.