Editorial Standards

1. Mission

Chatterbox exists to give residents clear, accurate awareness of what is happening in their immediate civic environment — their neighborhood, municipality, county, and school district. Our standard is clean local awareness: every item we surface must be true, sourced, and relevant to a real geography. We do not chase engagement metrics at the expense of accuracy, and we do not publish anything we cannot attribute to a verifiable source.

2. Sourcing Rules

• Source link required. Every item we display carries a link to the primary source document or page. We do not summarize claims that cannot be traced to a URL or document we have indexed.
• Two-source rule for factual claims. Any item that constitutes original reporting — meaning a factual claim not previously published by a source we index — must be supported by at least two independent, attributable sources before publication. Single-source factual claims are held in the review queue until a second source confirms or until we can verify the claim through official documentation.
• Partner content summarization. We summarize content from indexed local publishers only to the extent permitted by fair use and, where applicable, our content-sharing agreements. Summaries cite the original publisher by name with a link. We do not reproduce full articles.
• Official sources preferred. For government, public safety, school, and health items, we prefer primary official sources (agency websites, press releases, official document filings) over secondary coverage.

3. Provenance Labeling

Every item on Chatterbox carries exactly one provenance label indicating how it was sourced and whether AI assistance was used. The eight labels are:

• Original Reporting — Content reported and written directly by Chatterbox News editors, meeting the two-source standard. Dormant by default until an editorial team is in place.
• Official Summary — An AI-assisted summary of an official government or agency document (meeting minutes, budgets, inspection reports, etc.). The underlying document is the source of record.
• Official Notice — A verbatim or lightly reformatted notice issued by a government body, agency, or official office (permit notices, road closures, emergency alerts, etc.). No editorial interpretation is added.
• Local Media — A summary of or link to reporting published by an indexed local news outlet, within fair-use limits, with attribution and a link to the original.
• Community Submission — Content submitted by a verified community member or organization through the Chatterbox submission form. All community submissions are reviewed before publication.
• Event Listing — A community or civic event drawn from an indexed calendar feed or submitted by an organizer.
• Sponsored — Paid content from a sponsor. Clearly distinguished from editorial content; sponsor identity is disclosed. Sponsors have no editorial influence.
• AI-Assisted Summary — A summary generated by Claude AI from indexed source text. The underlying source is cited; the summary does not introduce facts not present in the source. See the AI Use Policy for full details.

4. Sensitive-Topic Rules

Items in the following categories are held in our review queue and are not shown in briefings or location pages until a human editor approves them. This is enforced at the server level — classification alone does not publish an item.

• Crime. Requires an official source (law enforcement press release, court filing, or agency report). We do not summarize crime reports from social media or unverified community posts. Victim identities are omitted unless officially disclosed and relevant.
• Arrests. We report arrests only when sourced from official booking records or agency statements. An arrest is not a conviction; items must reflect the presumption of innocence. We do not publish mug shots.
• Deaths. Obituary and fatality information requires official confirmation (medical examiner, agency statement, or verified family notification through an indexed publisher). We do not report unconfirmed deaths from social media.
• Minors. We do not identify minors by name in crime, arrest, disciplinary, or safety contexts. School achievement and event listings may include first names where the school or family has publicly disclosed them.
• Schools. School safety incidents, disciplinary actions, and allegations involving school staff require official source confirmation before publication. Routine school news (events, closures, achievement) follows standard sourcing rules.
• Elections. Election-related items — candidate announcements, results, election-administration notices — are sourced exclusively from official election-authority filings or verified candidate/party statements. We do not surface unverified election claims.
• Lawsuits. Legal filings are sourced from court dockets or official counsel statements. We summarize claims as alleged, not proven, and note the filing party. Items involving ongoing litigation carry this context.
• Public health. Disease outbreak, health advisory, and inspection items are sourced from official public health agencies (county health departments, CDC, state DOH). We do not amplify unverified health claims.

5. Social-Media Policy

Posts on social-media platforms (Facebook, X/Twitter, Nextdoor, Reddit, Instagram, etc.) are treated as tips, not publishable facts. We may use social-media content to identify a potential story for investigation, but we do not summarize, republish, or cite social-media posts as sources without independent verification through an official or indexed-publisher source. AI classifiers are explicitly prohibited from treating social-media content as factual input.

6. Originality and Attribution

We give credit for every piece of journalism and official information we draw from. Attribution appears in the item card (source name, link, provenance label). We do not present others' work as our own, do not strip author bylines, and do not rewrite partner content in ways that obscure its origin. Corrections to attributed content are coordinated with the original publisher where feasible.

7. Review-Queue Process

Items flagged as sensitive, uncertain, or community-submitted enter the review queue and are invisible to the public until an editor makes a disposition decision: approved, rejected, or escalated. Editors see the full classification rationale, the underlying source, and any AI confidence score. Approved items are published immediately; rejected items are archived. The review queue is the sole mechanism for releasing sensitive content — there is no bypass path.

8. Editor Responsibilities

Editors are responsible for: verifying sources before approving review-queue items; ensuring provenance labels match the actual content; confirming that AI summaries do not introduce facts not present in the underlying source; applying the two-source rule to original-reporting items; and publishing correction notices when errors are found. Editors may not approve items in categories where they have a direct personal or financial interest.

9. Conflicts of Interest

Staff and editors disclose financial, familial, and organizational relationships that could affect coverage decisions. Conflicted editors recuse from relevant review-queue items and correction decisions. We do not cover organizations in which an editor or staff member holds a significant financial stake without disclosure. Recusal and disclosure records are maintained in the audit log.

10. Independence from Sponsors

Sponsors have no editorial influence, directly or indirectly. Sponsorship agreements do not include any guarantee of coverage, positive or negative. Sponsor content is labeled Sponsored and is visually and structurally separated from editorial content. Editorial decisions — what to publish, what to hold, what to correct — are made without reference to sponsor relationships. A sponsor complaint about editorial coverage is treated the same as any other reader complaint.