Method & Standards

The Bright Line

Wiiver is human-judgment-first, AI-accelerated. We use software to gather, cluster, and organize the public record at a speed no individual could match, but the part that matters is always a person’s: the judgment. We state plainly which is which, because trust in analysis depends on knowing who is accountable for it.

What that looks like in practice. Take our SCOTUS closing-week piece. Software can retrieve both rulings, the one making regulators removable at will and the one striking party-spending caps, and file them under “law.” What it cannot do is judge that they are one story: that uncapped party money now funds the elections that swing the regulators the Court just unshielded, turning political risk into a standing input to regulatory forecasting. A person reasoned to that connection, decided it was sound, and signed it. That judgment, not the retrieval, is Wiiver.

How We Use AI

We use general-purpose AI assistants as research and production support. Specifically, they help us:

• Curate the News: scan and cluster developments across sectors and surface the ones worth a closer look;
• Do Deep Research: retrieve and organize primary sources, filings, dockets, disclosures, transcripts, releases, and assemble timelines;
• Rephrase Our Own Copy: tighten and clarify drafting a human has already written and owns; and
• Proofread: catch typos, inconsistencies, and unclear passages before a person makes the final call.

We treat everything an AI tool produces as unvetted raw material, not finished work. A person checks it, rewrites it, and takes responsibility for it.

What AI Does Not Do

• Decide what a story means. No forecast, causal claim, or judgment of significance is made on a tool’s authority.
• Publish. Every Analysis is reviewed and approved by a person before it goes live.
• Carry a byline. Every piece is authored and owned by a named human on the Wiiver Team.

How We Handle AI-Surfaced Facts

Our standing rule is that every source, quote, statistic, date, or image an AI surfaces is checked against the original before we rely on it, and we do not present AI-generated quotes or imagery as real. Verifying that is the reviewer’s core job. When an error slips through anyway, we correct it in the open (see Corrections and Updates).

Sourcing

Every Analysis is built on cited primary sources you can open and check. We separate what is established fact from what is analysis or forecast, and when we draw a conclusion we show the chain that led to it. When something is uncertain, we say so rather than rounding it up to certainty.

We treat the publishers and reporters whose work we build on as partners, not raw material. We link to original reporting, name the outlet, and quote only what is needed to make a point. The goal is to send you to the source, not to replace it.

Corrections and Updates

We will get things wrong, and when we do we fix them in the open, fully and promptly. We do not silently edit a published claim out of existence.

• Factual errors are corrected promptly and labeled clearly as corrections, with a short note of what changed and when.
• Developing stories are versioned: as facts firm up, dated updates are added rather than overwritten.
• To report an error, email corrections@wiiver.co.

Correction, August 27, 2026. This page previously carried a section titled “The Blindspot Score,” describing it as a feature some Analyses carry and stating that the name was “a plain description of what the measure does, not a trademark.” The component has been retired and the section has been removed. Both statements are withdrawn.

Independence and Conflicts

Wiiver’s analysis is not for sale. We currently run no advertising or sponsorship. If we introduce it, sponsored content will be clearly labeled and kept entirely separate from editorial judgment. Sponsors get no review of, or influence over, what we conclude, consistent with FTC guidance on native advertising and endorsements. Where a piece touches an entity we have any relationship with, we disclose it in the piece.

Right of Reply

When an Analysis makes a pointed claim about a named person, company, or agency, and especially where it could be read as alleging wrongdoing, we aim to reflect their side fairly. The more serious the criticism, the greater our effort to represent a relevant response, whether from the party, their public statements, or the record. If you are the subject of a piece and believe we got something wrong or left out your response, email corrections@wiiver.co and we will review it promptly.

Our Formats and Marks

Analysis, the Wiiver Weave™, and Wiiver’s signature sections (Key Context, Intersections, and Looking Forward) are part of how we work, and they are ours. The names Wiiver™, Wiiver Weekly™, and Wiiver Weave™ are trademarks of Wiiver, Inc., and the specific design and text of any given piece or Matrix are protected by copyright.

You’re welcome to share and cite our work, including exported Weave cards, as long as Wiiver’s branding and attribution stay intact and link back to the original. You can find the specifics in our Terms of Use.

Who Is Accountable

During the beta, every Analysis is produced and owned by a named human on the Wiiver Team, who reviews the sourcing, adopts or rewrites every claim, and approves publication. We are a small, independent operation based in Washington, D.C., writing for readers who work at the intersections we cover. To reach the person behind a piece, email corrections@wiiver.co.

Changes to These Standards

As Wiiver grows, these standards will too. We will post the revised version here and update the date above. If a change is material, meaning a shift in how we handle AI, sourcing, or corrections, we will say so plainly rather than quietly amend it.