AI Usage Disclosure

Orbis Signal briefings are produced through a fully automated, multi-stage AI pipeline. Artificial intelligence performs web-grounded research, analytical assessment, briefing composition, quality evaluation, and cross-domain synthesis. There is no human author, editor, or fact-checker in the production loop between scheduled pipeline runs and publication.

We disclose this openly because our readers deserve to know how their briefings are made. Automated production does not mean unaccountable production — editorial standards (2026-09-v3), source traceability, continuity discipline, and quality controls are enforced architecturally and through validation code.

Each edition is produced on a scheduled daily pipeline. Production is organised around clear boundaries rather than a single free-form generation step:

• Framing — establishes the editorial lens for the edition
• Research — gathers claims and tier-labelled sources from live public reporting into a locked evidence record
• Analysis — derives judgments from that evidence with epistemic labels and confidence bands
• Composition — renders the structured public briefing
• Quality control — structural checks and scorecard evaluation before publication
• Cross-domain synthesis — optional macro-theme analysis when all topic briefings clear quality control

Continuity standards apply across editions: recurring storylines are expected to lead with material change rather than restating prior baseline coverage.

The pipeline uses models from two providers. Exact model versions may change as capabilities evolve; the editorial policy version, not the model version, is the primary traceability marker for published briefings.

• Google — web-grounded research and edition framing (live search retrieval)
• OpenAI — analytical assessment, briefing composition, quality evaluation, and cross-domain synthesis

After the evidence record is locked, analytical and writing stages do not have access to the open web. They work exclusively from recorded claims and sources, which prevents unverified links from entering composition.

What humans do:

• Design, maintain, and update editorial standards, validation rules, and the quality scorecard
• Monitor pipeline health, investigate failures, and respond to reader correction requests
• Operate the publication schedule that moves validated briefings from production to the public site

What humans do not do:

• Write, edit, or approve individual briefing text before publication
• Manually verify each source URL or claim in the production loop
• Hand-pick which developments appear in a given edition
• Override quality control failures to publish substandard briefings

AI-generated content is constrained by structured schemas, cross-stage validators, and editorial standards enforced in code:

• Every output must conform to a typed schema — free-form generation without structure is not permitted
• Source URLs in published briefings must exist in the locked evidence record
• Key judgments must carry epistemic types and confidence bands
• Sensationalist phrasing and AI disclaimer language are rejected
• An eight-dimension quality scorecard — including continuity — must pass before publication

These safeguards are described in detail in our Verification Policy.

AI models can hallucinate, misinterpret sources, or produce overconfident language despite safeguards. Our architecture mitigates these risks through evidence locking, URL validation, epistemic labelling, and continuity standards — but no automated system eliminates error entirely.

Readers should treat Orbis Signal briefings as structured analytical products produced by AI under defined editorial standards, not as human-authored journalism. We encourage critical reading and welcome error reports under our Corrections Policy.

We will update this disclosure when we materially change our AI providers, add or remove major production stages, or introduce human review steps. Model version changes that do not alter the production architecture may not trigger a disclosure update.

Related: Editorial Policy, Methodology, Philosophy.