Skip to main content
Glossary · Definition

Fact-checking

Fact-checking is the process of verifying a claim against reliable evidence and publishing a finding about whether it is accurate, misleading or false.

Last updated September 28, 2026

How fact-checking works

A checker starts by choosing a claim that is specific, checkable and in public circulation. Opinions and predictions are generally not checkable. The checker then traces the claim to its origin, looks for primary sources such as documents, data, court records or direct statements, consults subject experts, and gives the source of the claim a chance to respond. The result is a published rating or explanation with the evidence linked.

Practices differ between organizations, and several fact-checking groups publish their methods and rating scales. Some platforms partner with independent checkers, and some use labels or community notes written by users. The labeling approach varies by platform and changes over time, so a team should check current policies rather than assume them.

Fact-checking inside an organization

Comms teams do their own version. Before responding to a claim, they establish what is true: what the internal records show, what a reasonable outsider could verify, and which parts of the claim are accurate. A response that concedes the accurate part and corrects the false part is more credible than a blanket denial. Keep records of the sources used and the time of the check, since the situation can change.

Why it matters to communications and risk teams

An independent fact-check can be the most durable reference for a disputed claim, and reporters and platform reviewers look for it. It can also be slow. A claim can complete most of its spread before a checker publishes. That gap is one reason teams track claims from the first appearance rather than waiting for a rating.

What a good rating record contains

A published check that can be relied on states the claim as it was made, where and when it appeared, the evidence reviewed, and the reasoning for the rating. Teams that cite a fact-check in a response should link the original and quote it accurately, rather than paraphrase the rating. They should also note the date, since ratings can be revised when new evidence appears.

Speed and scope

A fact-check takes time, and claims often move faster. Teams can use a staged approach: confirm the core facts internally within hours, publish a short factual statement, and then update it as more detail is verified. Noting what is still being checked is honest and avoids overstating certainty. Correct promptly when new evidence changes the picture.

Common misconceptions

A fact-check is not a rebuttal service for hire. Credible checkers decide what to review and do not take instructions from the subject. Requesting a review is fine; expecting a favorable outcome is not. Fact-checks also do not remove content by themselves. What happens to a rated post depends on each platform's rules.

Finally, checking a claim is different from assessing its source. A reliable outlet can be wrong once, and a low-quality source can be right once. See source credibility.

See how these signals show up in your own coverage on the PeakMetrics platform or run the free AI Perceptions check. Back to the glossary.

Track these signals in your own coverage.

PeakMetrics follows news, social and broadcast sources, and what AI assistants say about your organization.