Research notes where every AI action is on the record.
Outline, draft, and review with AI assistance, then publish articles with a built-in provenance ledger so readers can see how the work was made.
Transparent AI Research Notes
A practical note on preserving trust when AI systems help researchers outline, draft, review, and publish evidence-backed writing.
2. The provenance ledger
A provenance ledger is the smallest unit of disclosure a research surface should carry. For each AI action we record provider, model, purpose, and prompt_summary.
Disclosure is not an apology for AI use. It is a record of how the work was made.
Create a structured research outline with citation checkpoints.
Expand section 2 with three paragraphs and retain placeholder citations.
Surface methodology risks and citation gaps before human review.
Three things every article carries.
AI assistance is allowed. Disclosure is not optional. Provenance shows up wherever the article does.
Provider, model, purpose, and prompt summary are recorded for every AI run.
Verified, needs-review, and missing states are visible on every surface.
AI peer review and human review pass-through are tracked separately.
Outline to draft to review to publish.
Markdown writing
Plain markdown body with abstract, status, type, and version.
AI provenance
Outline and draft tools log every action to the ledger.
Citation checks
Verified, needs-review, and missing states on every citation.
AI peer review
Strengths, risks, gaps, and methodology questions stay separate from text.
Version history
Snapshots with status, summary, and timestamp.