How f/seek Fact-Checks a Political Claim

Every day, politicians make hundreds of statements — in hearings, press conferences, social media, and on the floor of Congress. Most go unverified. Some are true. Some are false. Many live somewhere in between, stripped of the context that determines whether they hold up.

f/seek is built to change that. Here's exactly what happens when a political claim enters our platform.

Step 1 — Anyone Can Submit

Verification starts with a user. Anyone can submit a political claim — a quote, a statistic, a policy assertion. The only requirement: the claim must be specific and attributable. Vague opinions don't qualify. Checkable statements do.

This open submission model means f/seek's corpus is driven by what citizens actually encounter and want verified — not what editors choose to cover.

Step 2 — Source First, Always

Before any analysis begins, a source must be attached. This is non-negotiable. The claim must be anchored to something verifiable: a congressional record, a CBO report, a published study, a video timestamp.

If no source is attached, the claim stays in draft. This single gate eliminates the vast majority of low-quality submissions before they ever reach the verification layer.

Step 3 — AI-Assisted Analysis

Once sourced, the claim goes through structured AI analysis using our verification pipeline. The model cross-references the claim against web sources, official databases, and prior verified claims on the platform.

The output is not a simple true/false — it's a structured breakdown:

• →Verified True — the claim is accurate and the source confirms it
• →Mostly Correct — accurate in substance but missing context or precision
• →Needs Context — technically accurate but misleading without more information
• →Mostly False — significant inaccuracies in the core assertion
• →False — the claim is not supported by credible evidence
• →Unverified — sources were insufficient to reach a verdict

This tiered system reflects reality. Political claims rarely live in clean true/false territory.

Step 4 — Community Review

AI analysis is a starting point, not the final word. After the initial verdict, the claim enters community review. Other users can add follow-up sources, challenge the reasoning, or flag context the model missed.

This layer is what separates f/seek from a simple fact-checking API. The community doesn't vote on verdicts — they contribute evidence. The quality gate ensures only sourced follow-ups move the needle.

Step 5 — Verdict Published with Structured Data

Once a claim passes the quality threshold, it's published with machine-readable structured data attached using JSON-LD — the same standard Google uses to understand and surface content in search results.

Every piece of content on f/seek is annotated with schema that describes exactly what it is:

• →Congress member profiles use Person schema — name, role, chamber, and legislative context
• →Bill pages use Legislation schema — bill number, sponsor, status, and summary
• →Blog posts like this one use BlogPosting schema — headline, author, date, and publisher
• →All pages carry WebPage schema with canonical URLs and site context

This structured layer means search engines don't have to guess what f/seek content is about — it's declared explicitly in every page. A senator profile isn't just text about a person; it's machine-readable data about a sitting lawmaker. A bill page isn't just a summary; it's a structured legislative record.

That precision is what makes f/seek content durable and discoverable, not just readable.

Why the Process Matters

The internet has no shortage of hot takes. What it lacks is structured, verifiable records of what was said, sourced, and checked.

f/seek's five-step process is designed to produce claims that are durable — worth indexing, worth sharing, and worth building a public record on. Every verified claim is a small act of civic infrastructure.

"The standard you walk past is the standard you accept."

We're not walking past unverified political statements anymore.

Ready to submit your first claim? Open the app and start fact-checking.