Trust & Transparency

The value we offer is not a promise of flawless objectivity — no newsroom can honestly make that promise. Our value is a transparent process: a clear standard for what we cover, an honest account of how the work is made, and a public record of every time we get something wrong. This page lays out that process in plain terms. It is meant to be checked against what we actually publish.

Two commitments sit under everything below. The first is the human consequence of the news — who is affected, and how their daily life changes. The second is factual clarity that can lower the temperature of public life rather than raise it. Clarity is not the same as splitting the difference between two sides; it is stating what the record shows, plainly, so readers argue about what to do instead of what is true.

How We Choose Stories

A story earns a place in the edition when it meets a consistent standard, applied every day across every section:

• Consequence and public impact. The story changes something for real people — their money, health, safety, rights, schools, or communities. We ask "so what?" before "what happened?"
• Timeliness. It matters now, either because it is developing or because a decision, deadline, or ruling gives it fresh weight.
• Broad relevance. It reaches beyond a single niche or interest group and speaks to a wide readership.
• Verifiability. The core facts can be confirmed against primary documents, data, or direct statements — not just against other coverage.
• Importance across communities. We weigh how a story lands across race, party, region, generation, and class, and we center the people living it rather than treating whole communities as an afterthought.

What We Leave Out

Just as important as what we run is what we decline to run. We deliberately exclude:

• Outrage bait. Items engineered to provoke anger rather than inform.
• Unsupported claims. Assertions we cannot tie to a credible primary or verifiable source.
• Repetitive updates. Incremental restatements of a story we have already covered, with nothing materially new.
• Virality for its own sake. Stories chosen only because they are trending, absent real consequence.
• Empty items. Coverage that adds no meaningful understanding — noise dressed as news.

Human Review and the Role of AI

We use AI, and we are direct about how. AI tools assist with research, curation, and drafting — scanning wide source sets, surfacing candidate stories, and preparing first drafts. AI does not have the final word.

Every edition is reviewed by a human editor who approves, edits, or skips each piece before it is published. The editor is accountable for what runs. The standards we place on the AI are explicit and non-negotiable:

• It must not invent facts, quotes, statistics, scores, votes, or sources.
• It must not fabricate section names, bylines, or attributions.
• Where the record is uncertain, it must say so rather than fill the gap.

When those lines are crossed in drafting, the human editor's job is to catch it and correct it before publication — and, if something slips through, to correct it in the open afterward. See Corrections, below.

Corrections

When we get something substantive wrong, we fix it visibly — never silently.

• On the original page. A substantive error is corrected on the story where it appeared, with a timestamp and a plain explanation of what changed and why.
• In the next edition. The correction also appears prominently at the top of the next edition, set in type comparable to the original error — not buried.
• In a permanent public log. Every published correction is recorded in our Corrections Log, which stays public and does not disappear.
• No silent corrections. We do not quietly alter a published story to erase a mistake. The record is the record, and the correction is part of it.

See the full, running record in the Corrections Log →

Sources & Receipts

Our verification feature, Receipts, and our sourcing across the edition follow one rule: get as close to the record as possible.

• Primary sources first. We prefer primary documents, official data, and direct statements — roll-call votes, court rulings, agency releases, filings, and full quotes — over second-hand characterizations.
• Clear labeling. We distinguish what is confirmed, what is disputed or missing, and what is unknown. We do not present a disputed claim as settled, or a settled fact as merely one opinion.
• Permanent links. Where possible, we link directly to the primary source so readers can check the record themselves.