Skip to content
NEWSMESIS

AI-assisted, human-accountable newsroom

Trust & accountability

Newsmesis uses AI to accelerate collection, clustering, drafting, and formatting. Humans remain accountable for publication, corrections, and high-risk editorial judgment.

Publication controls

  • Two-person approval before publication
  • Verification ledger for material and high-risk claims
  • Embargo and scheduled-publication enforcement on reader surfaces
  • Visible corrections with downstream propagation status

Source licensing matrix

articleFull source links, excerpts, and correction notices require cleared or public-interest source rights.
cardsCard exports require rights review for text, image, logo, and agency material before distribution.
shortsShort-form scripts require explicit clearance when source material is transformed for social video.
rss_apiMachine-readable feeds carry only cleared summaries, canonical URLs, and correction metadata.

Corrections

Corrections and retractions are preserved in article update logs and the public corrections archive. Downstream surfaces are tracked until completed or marked not required.

High-risk review

Stories involving named people, markets, elections, investigations, medical claims, or defamation risk require extra standards review and right-of-reply status.

Takedown and rights concerns

For rights, safety, or takedown concerns in this local prototype, contact takedown@newsmesis.local.

Current implementation is local-first and no-secret: data lives in the repository data store, and production identity, legal workflow, and hosted database controls remain planned work.