Automation transparency

Automation may assist; it cannot approve

Automation may monitor allowlisted sources, identify possible developments, organize source packets, compare specifications, draft text, suggest links, and run policy or quality checks. Those uses can improve speed and consistency, but they do not create editorial authority.

A real human must review the evidence, claims, originality, identities, image rights, disclosures, and final presentation before an article can become public. A material change invalidates the exact approval bound to the article and its ledger.

Prohibited uses

Automation may not:

• publish or send a newsletter without the required human action;
• invent a source, quote, test, interview, byline, reviewer, expertise, or first-hand experience;
• treat a search snippet, AI summary, or another outlet’s unsourced claim as evidence;
• bypass a blocked claim, missing right, stale approval, or high-risk review requirement;
• mass-produce near-duplicate pages for search volume; or
• hide a correction or overwrite the public history.

Reader-facing disclosure

An article carries a specific automation disclosure when automated assistance materially affects reader trust. Boilerplate is not a substitute for explaining the meaningful use. Organizational automation identities are never presented as people.

The private editorial ledger preserves more detailed source, claim, rights, review, correction, and approval records; it is not copied into the public site. Read the editorial standards for the complete gate and Contact for questions.