TECHNOLOGY COMPANIES under fire for refusing to fact-check political advertisements on their platforms are pursuing all manner of responses — except fact-checking political advertisements on their platforms. That’s a mistake.

Free speech has always been a foundational value in Silicon Valley, and leaders such as Mark Zuckerberg are right to champion it against the corrosive censorship model on offer in China and elsewhere. Politicians should, for the most part, be able to lie on Facebook, just as anyone else is, and the public should be able to hold leaders to account. But that’s a different question from whether politicians should be able to pay to have their lies spread, based on unprecedentedly precise behavioral data, to the voters who are most likely to believe their lies.

Twitter’s response to this quandary has been to ban political ads, full stop. Or so it thought; it turns out that defining what’s political is almost as difficult as defining what’s true. Twitter caught flak almost immediately for the possibility that it might permit ExxonMobil’s elegies to the oil industry while barring green groups’ climate ads. The move is also a boon to incumbents and a burden for upstarts clambering to build a list of donors, or to catch the public’s notice at all.

Google’s reply has been more nuanced. The company will limit the criteria campaigns can use to “microtarget” ads to narrow audiences based on party affiliation or voter record. The aim is to increase accountability by letting more people see ads and to encourage messaging that seeks to persuade rather than push people further into preselected echo chambers. Yet this could pose the same problems for grass-roots organizers as a broader ban by producing less bang for each fundraising buck. Microtargeting can also be positive, drawing the disengaged into the conversation by showing them something relevant: a veteran who wants to know what a candidate will do for her health care, or an immigrant who has so far stayed on the sidelines.

Then there’s Facebook, which, at least for now, is doing nothing except making political ads (insufficiently) searchable in an archive. Mr. Zuckerberg has said it shouldn’t be his job to play democracy policeman. He’s not wrong. There’s room for Congress and the Federal Election Commission alike to act on digital advertising, especially when it comes to imposing standardized transparency requirements. But Facebook has become so powerful that doing nothing has as much of an impact as doing something — and in this case, a negative one.

Facebook should err toward permissiveness in assessing ads for truthfulness, especially when it comes to issue ads. The Republican Party, for example, probably should be allowed to say climate change is a hoax, even though all scientific evidence says otherwise. Voters have an interest in knowing whether a party has descended to the know-nothing basement. But the platform’s fact-checkers should know a blatant misrepresentation when they see one. If an opponent wants to pay to say Barack Obama was born in Kenya, Facebook should just say no.

Read more: