ClaimReview: Fact Check Schema That Signals Trust to AI and Search in 2026

ClaimReview is a piece of structured data, defined by Schema.org, that publishers add to a fact checking article to describe what claim was assessed, who originally made it, and what the verdict was. In plain terms, it is a machine readable label that says: here is a claim, here is our evaluation, and here is how true we found it. Search engines, fact check databases, and increasingly AI systems read this markup to understand the credibility of a statement.

For marketers, journalists, and anyone publishing in trust sensitive areas, ClaimReview matters because it converts a written fact check into data that engines can interpret. As discovery moves toward AI answers that must judge reliability, structured trust signals like this one help machines tell a verified claim from an unverified one.

What is ClaimReview?

ClaimReview is the Schema.org type used to mark up a fact checking review of a claim made or reported in some other work. A single block of ClaimReview markup captures the essentials of a fact check: the claim under review, the rating that summarizes the verdict, the URL of the full review, and the organization or person behind the assessment. It is one specific application of structured content aimed squarely at verification.

The point is to remove ambiguity. A human reader can tell that an article is debunking a rumor, but a machine needs explicit fields to know which statement was checked and whether it was judged true or false. ClaimReview provides exactly those fields, which is why it underpins dedicated fact checking products rather than ordinary article markup.

How ClaimReview markup works

You embed ClaimReview as JSON-LD, Microdata, or RDFa on the page that hosts your fact check. The markup names the claim, attaches a verdict, links to the full review, and identifies the publisher. Engines and fact check tools then parse that data and can present a condensed version of your assessment to people searching for the underlying claim. Done well, it forms a clear bridge between your conclusion and a machine that needs to act on it, a core goal of AI grounding.

For eligibility in fact check features, the data must match the visible page content, the URL must sit on the same domain or subdomain as the hosting page, and the publisher must follow the relevant guidelines. Google also notes that to qualify for a single fact check result, a page should contain only one ClaimReview element, which keeps each assessment unambiguous.

Key properties of ClaimReview

Three properties are required. The first, claimReviewed, holds a concise summary of the assessed claim, ideally under 75 characters, and should not contain the rating itself. The second, reviewRating, carries the verdict and supports both a numeric value and a textual label, with the text shown to users. The third, url, points to the full fact check and must live on the same domain or subdomain as the page.

Two properties are strongly recommended. The author property names the publisher of the fact check, an organization or person, ideally with a name and a URL. The itemReviewed property, modeled as a Claim, describes the original statement, including who made it and when it appeared. Naming the original source clearly is essential, and it connects to careful source citation across the article.

The ClaimReview rating scale

ClaimReview uses a standardized numeric scale from 1 to 5 to summarize a verdict. A value of 1 maps to False, 2 to Mostly false, 3 to Half true, 4 to Mostly true, and 5 to True. The numeric value gives machines a consistent signal, while the accompanying text label is what readers actually see in a result.

This consistency is what lets independent fact checkers around the world contribute to shared databases that can be queried at scale. A common scale means an AI system can compare verdicts across publishers without having to interpret each outlet's bespoke wording, which strengthens reliable multi-source synthesis.

ClaimReview, Fact Check Explorer, and the markup tools

ClaimReview powers a set of dedicated tools. Fact Check Explorer lets anyone search claims by keyword and see matching fact checks from reputable publishers, filterable by language and source. The Fact Check Markup Tool lets publishers submit ClaimReview data through web forms rather than hand coding it, and that submitted markup is treated the same as markup embedded directly in the article. An accompanying API lets authorized webmasters create and edit markup programmatically through Search Console.

Crucially, Google has been phasing out ClaimReview rich results in standard web Search, but the markup remains fully supported by Fact Check Explorer. So while the SERP display has narrowed, the structured data still feeds the fact checking ecosystem and the databases that many tools, and likely AI systems, draw on.

Why ClaimReview matters for SEO and GEO

Even with rich results being phased out, ClaimReview remains a credibility signal. It explicitly demonstrates the rigor behind your content: a named claim, a traceable methodology, citations, and a clear verdict. That rigor aligns with the experience, expertise, authoritativeness, and trust principles captured by E-A-T, which search and AI systems lean on when judging sources.

For generative engine optimization, the value is about being trusted at the claim level. AI answers must avoid repeating false statements, so machine readable verdicts give them a reliable signal about what holds up. Publishing verifiable, well sourced assessments helps reduce the spread of AI hallucination and positions your domain as a source worth grounding answers on.

How to implement ClaimReview correctly

Start by writing a genuine fact check that clearly identifies the external claim, presents your conclusion, and shows the reasoning with citations and primary sources. Then add ClaimReview markup whose fields exactly mirror the visible content, with no mismatch between data and page. Validate the markup before publishing to catch missing required properties.

Meet the eligibility expectations too. Publishers generally need multiple marked pages, a corrections policy or error reporting mechanism, and clear attribution of each claim to its external origin. Sites belonging to political entities are not eligible. Treating ClaimReview as part of a broader, well organized content system, supported by sound keyword research and content planning, keeps your fact checks discoverable and consistent.

Challenges and limitations

The biggest limitation is that markup never guarantees display. Structured data makes a feature possible, not certain, and with rich results being phased out in Search, the most visible payoff has shifted toward Fact Check Explorer and the wider database ecosystem. Publishers should set expectations accordingly rather than assuming a SERP badge.

Implementation is also unforgiving. Mismatches between markup and page content, missing required properties, or off domain URLs can disqualify a page. Maintaining ClaimReview at scale demands editorial discipline and ongoing validation, which is why it tends to be used by dedicated fact checking operations rather than casual publishers. It pairs naturally with strong author authority.

Conclusion

ClaimReview turns a fact check into structured data: a named claim, a 1 to 5 verdict, a link to the full review, and the publisher behind it. Although Google is phasing out the rich result in standard Search, the markup still powers Fact Check Explorer and the broader fact checking ecosystem, and it remains a meaningful trust signal for both search and AI driven discovery.

For trust sensitive publishers, it is a way to make rigor machine readable. Combine it with strong E-A-T signals and clean structured content to reinforce credibility. Reference sources: Google Search Central, IDX, and Schema.org.