Mission

Why Veridi exists

Most people evaluate claims the same way: they read something, decide whether it sounds right, and move on. Even professional fact-checkers often rely on experience and instinct to guide their assessments. This works until it doesn’t - and in an environment where disinformation is engineered to exploit exactly these shortcuts, it increasingly doesn’t.

Veridi takes a different approach. It provides an explicit, step-by-step process for evaluating any factual claim, regardless of subject matter. Think of it as a checklist; not for what to believe, but for how to evaluate.

What we’re building

Veridi is a structured fact-checking methodology, implemented as a prompt system for AI (currently Claude by Anthropic), which follows the documented procedures to evaluate claims. The AI doesn’t decide what’s true; it follows the methodology’s decision trees, source hierarchy, and quality checks to produce a transparent, auditable assessment.

This matters because the methodology is separable from any particular AI system. The procedures are documented, the decision logic is explicit, and the assessment criteria are inspectable. If the AI makes a mistake, the methodology provides a framework for identifying where and why.

Principles

Truth comes first. Every assessment begins by stating what is actually true, clearly, affirmatively, and without negation. The claim under review appears second. This ordering exists because psychological research shows that repeating a false claim, even to debunk it, can reinforce it. Veridi leads with the corrective.

Transparency over authority. We don’t ask anyone to trust us. We publish our methodology, our test results, our limitations, and our reasoning. If our process is sound, the results will bear that out. If it isn’t, we want to know.

Forthright uncertainty. When the evidence is ambiguous, we say so. When experts genuinely disagree, we report the disagreement rather than picking a side. Veridi uses nine verdict categories, not because nuance is fashionable, but because binary true/false fails to describe most real-world claims accurately.

Adversarial awareness. Disinformation is not random. It follows patterns: techniques that exploit specific weaknesses in how people and institutions evaluate information. A fact-checking system that doesn’t account for this is incomplete. Veridi identifies eleven such patterns and provides explicit detection procedures for each.

Organization

Veridi is being built by a small Canadian organization with intent to establish as a nonprofit. There is no venture capital, no growth-at-all-costs pressure, and no advertising model. The organization exists to produce accurate fact-checks and to make its methodology available to others.

We are not competitors to existing fact-checkers. Organizations like PolitiFact, Snopes, and the IFCN network do important work, and prior art such as Loki (an open-source fact-checking platform) brings value and identifies challenges. Veridi’s contribution is a structured, auditable methodology that can complement and strengthen existing practices - not replace them.