SEN. ELIZABETH WARREN (Mass.) was right during Tuesday’s Democratic debate to resist Sen. Kamala D. Harris’s (Calif.) entreaty that she join in demanding President Trump’s removal from Twitter. But another question about another platform’s role in policing politicians is trickier.

Ms. Harris thinks Twitter should suspend the commander in chief for the simple reason that he is violating its rules. It’s a bad idea for the same reason platforms are smart as a general matter to avoid meddling in the democratic discourse. A private company picking which politicians are permitted to communicate with the public is a dubious proposition, and society generally benefits from citizens seeing the speech of those who represent them. The more information available, the more informed the voters.

But what happens when that information is false?

This is the morass Facebook has been stumbling through as it has attempted to articulate its attitude toward political advertising. The company argues that the content politicians pay to have promoted shouldn’t be subject to fact-checking even though its advertising policies prohibit falsehoods from anyone else. That excuses content such as a recent attack ad from the president’s reelection campaign full of widely debunked claims about Hunter Biden’s dealings in Ukraine.

Facebook, in this case, isn’t determining whether Mr. Trump or any other politician has the right to speak at all; it’s determining whether it will accept money to spread deliberate falsehoods across its platform. Millions of people have viewed the bogus Burisma ad. The ability to target narrow audiences amplifies the problem, and so do Facebook’s viral mechanics. Candidates can carefully reach low-information voters en masse. The media may refute an offending ad, but unsensational honesty will not travel as far as sensational deception does with the help of Facebook’s engagement algorithms.

Equal-opportunity requirements for broadcasters require them to allow latitude to lies, but cable networks can reject ads that contain flat-out falsehoods. Some did exactly that last week, and Facebook should, too. These decisions won’t always be clear-cut. There’s the concern that Facebook could play games — or get played by the conservatives hounding the platform over nonexistent censorship. The electoral arena has also always played host to hyperbole, and locating the line between permissible exaggeration and unacceptable smear will require some thought. But responsible thinking, plus a hefty helping of transparency and a robust appeals process, is exactly what’s necessary.

Mr. Trump and other conspiracy-mongers might chafe at their inability to poison the public conversation with such precision at such scale. They can take comfort, perhaps, that their Twitter accounts probably aren’t going anywhere anytime soon.

Read more: