A recent Washington Post headline claimed its tech columnist, Geoffrey Fowler, had shown that “Meta’s new crowdsourced system to fight falsehoods [has] failed to make a dent.” The claim would fail a proper fact check.
Meta launched its new program—Community Notes—on April 7 to replace third-party fact-checking. If you took the Post’s headline at face value, you’d expect Fowler to have measured misinformation levels before and after the launch. He didn’t; instead, he sent Meta 65 “debunking” notes on conspiracy theories, deepfakes, and hoaxes. Only three were published. From this, the Post concluded Meta’s system was a failure.
This is not fact-checking. It’s an opinion disguised as a finding.
While fact-checking organizations date back to at least the 1990s, their numbers surged during Donald Trump’s first presidency: from 176 in 2016 to 449 by the time he left office, now sitting at 442. This growth coincided with Facebook (now Meta) funding more than 45 percent of the industry’s revenue as of 2024, which may be going away.
Yet despite all the growth—or maybe because of it—many fact checkers treat their beliefs as facts.
The International Fact-Checking Network’s Code of Principles and the College of Staten Island’s fact-checker guide offer guidance to fact checkers, but never define the term fact. In disciplines such as science and philosophy, a fact is something that can be proved or disproved with objective evidence. Opinions, however well informed, are not facts.
Consider FactCheck.org’s August 4, 2025 headline: “No Evidence for Trump’s Claims of ‘Rigged’ or ‘Phony’ Job Numbers.” That is not provable. No human—or AI—can scour all existing information to prove no evidence exists anywhere. The Post’s retiring fact checker Glen Kessler was also prone to making no-evidence claims. These are opinions, not fact checks.
Or take FactCheck.org’s July 29, 2025 entry: “Trump’s Misleading Justification for Higher Tariffs on Imports of EU Goods.” To mislead means to lead in a wrong direction. FactCheck.org’s proof that Trump was misleading consisted of a single economist’s remark and vague references to unnamed economic experts. Similarly, a Kessler fact check of Senator Cruz on COVID origins was based on “interviews with actual scientists.” These aren’t proofs; they’re appeals to authority.
Some outlets do better. Politifact often focuses on statements that can be verified or disproved, but even it strays. For example, it assessed Trump’s July 28 comment, “I’ve stopped six wars — I’m averaging about a war a month.” Absent an objective and definitive counterfactual, the “mostly false” assessment is simply a difference of opinion on the evidence.
The issue isn’t semantics. When fact checkers wade into opinion arbitration, fact-checking morphs into political commentary cloaked as impartiality; and when people rely on fact checkers for a clear line between what’s demonstrably true and what’s merely disputed, they are being misled. When readers discover the deception, they lose trust in fact checkers. No wonder people believe that Republicans and Democrats cannot agree on basic facts!
This isn’t hard to fix. First, fact checking should stick to claims that can be objectively tested. What the government says is the federal budget surplus can be checked—whether Trump’s statements about it are hollow cannot. Second, fact checkers must be candid about the limits of what can be known. Admitting that something cannot be tested builds more credibility than stretching for a verdict. Third, they should stop laundering expert opinion into fact. Economists, scientists, and analysts often disagree. Citing one side as establishing fact is advocacy, not verification.
Meta’s Community Notes may or may not be effective. Measuring that would require careful before-and-after comparisons, large-scale content sampling, and clear criteria for misinformation. The Post’s method—send in notes, then object when they aren’t published—doesn’t meet that bar. The headline writers should have known better.
There’s a role for genuine fact-checking. When done correctly, it anchors public debate to a shared reality. When done incorrectly, it creates the very misinformation it’s supposed to counter. If fact checkers want to keep their title, they should return to the craft: testing fact claims against verifiable evidence. If they’d rather referee opinion disputes, they should drop the pretense and call themselves columnists.
That’s my opinion. Not a fact.
The post When Fact Checkers Stop Checking Facts appeared first on American Enterprise Institute – AEI.
