Useful claims need visible boundaries.
This policy explains who publishes Fluxreading content, what evidence can support a claim, how AI is used, and what happens when a page needs correction.
Observation, interpretation, and verification remain separate.
Every guide and case study should make clear what was directly observed, what was inferred through the published methodology, and what requires an on-site measurement, official source, inspection, or qualified professional.
Authorship and accountability
Current guides are published under the Fluxreading editorial team shared byline. We use a team byline because named contributors and specialist credentials have not yet been publicly verified. A shared byline does not reduce accountability: publication dates, material update dates, sources, and the responsible review standard appear on each article.
If named practitioners or technical specialists contribute later, their role, relevant qualifications, and scope of review will be stated without implying expertise outside that scope.
Research and source selection
We prefer first-party technical documentation, authoritative public datasets, and clearly identified internal methodology documents. Sources are linked near the claims they support. Traditional Form School interpretation is identified as interpretation rather than presented as an engineering or scientific finding.
- Map and elevation evidence is described at the scale it can support.
- Missing evidence lowers confidence; it does not become a negative fact.
- Safety, hazard, legal, financial, structural, and environmental claims require appropriate authoritative or professional sources.
- A material claim should be removed or qualified when its source cannot be checked.
Review and use of AI
Editorial drafts and diagrams may be assisted by software, including AI tools. Software assistance is not listed as an author and does not replace editorial responsibility. Before publication, content is checked for consistency with the deterministic rulebook, unsupported certainty, duplicated scoring concepts, source quality, privacy, and potentially harmful practical claims.
AI-generated property observations in the product are passed through deterministic scoring and evidence gates. The editorial team does not treat model output as a verified physical fact merely because it is fluent or specific.
Case-study labels and consent
Every case is labeled either Illustrative composite or Verified anonymous case. Composites teach a decision pattern and must not imply a real client, address, or testimonial. A verified anonymous case requires separate publication consent, source review, deliberate de-identification, and a final privacy check. A saved or shared reading is never publication consent.
Updates and corrections
Routine copy edits do not change the displayed review date. Changes to conclusions, evidence, methods, sources, disclosures, or practical guidance do. Material corrections are added to a visible correction note on the affected page, with the date and a concise description of what changed.
Correction requests can be evaluated when they identify the page or share link, disputed statement, and supporting evidence. A request does not make submitted material independently verified; changes are recorded according to the evidence and review standard above.
A public corrections address must be configured before broad commercial launch. Until then, owners should revoke disputed share links and the site should not publish real-address case studies.
Editorial independence and conflicts
Current editorial pages contain no paid placement, affiliate ranking, or sponsored property recommendation. If commercial relationships are introduced, disclosures must appear on the affected page before publication. Payment, partnership, or access must not determine the score or suppress a material evidence limitation.