AI Disclosure
Readers deserve to know how the content they read is made. This page explains, plainly, how artificial intelligence is and is not used at The Classic Note.
The short version
Articles on The Classic Note are written with AI and published through an editorial process that verifies facts automatically before anything goes live. A human editor sets the standards that process enforces, supervises and audits it, and handles everything it flags — but does not hand-edit every page before publication.
What we do not do
- We do not publish raw, single-pass AI text. No article is published unless it has passed the review and fact-checking steps described below.
- We do not invent quotations, sources, or biographical details. If a claim cannot be supported, it does not run.
- We do not pretend a human has read every page before it goes live. Human oversight here works at the system level — standards, verification, auditing, and corrections — not as a page-by-page sign-off.
- We do not backdate. Article dates reflect when a page was first published; when a page is substantially revised, its modified date changes.
How an article is made
- Drafting. AI writes the draft — research, structure, and prose — following editorial standards (scope, tone, structure, sourcing rules) defined by a human editor. This includes listening notes and recording recommendations, which are drafted within the same standards: treat them as a starting point for your own listening, not a definitive critical judgement.
- Automated review. Each draft goes through independent checks: a critical review for factual red flags, readability, and structure, plus validation of images, embedded videos, and recording details.
- Fact-checking. Verifiable claims — dates, catalogue numbers, premieres, quotations, historical events — are logged with a supporting source and checked before publication. High-risk claims are re-verified by a separate automated check that searches live web sources. If a claim cannot be supported, publication is blocked.
- Publication. Only drafts that clear these checks are published. Drafts that fail are held for human review or discarded.
What stays human
- Editorial standards. What we cover, what a good listening guide looks like, what counts as a reliable source, and what is never allowed — invented quotes, unsourced claims, hype — are human decisions, encoded into the automated checks described above.
- Supervision and auditing. A human designs the process, audits its output, reviews articles the checks flag or hold, and strengthens the checks when something slips through.
- Corrections. Error reports are read and resolved by a person, and every published change can be rolled back to its prior version.
Accuracy and corrections
No automated system is perfect. When an error slips through, we fix it and note substantive corrections. If you spot something wrong, please tell us on our Corrections page or through Contact.
We do not name specific AI tools on this page, as the tools we use may change over time; the process and standards described here are what stay constant.