This is a sample article created to demonstrate the platform. It is not real research: the text, citation states, and AI usage records are demonstration data.

PublishedNeeds reviewResearch note0.3.0

Transparent AI usage in research note workflows

A practical note on preserving trust when AI systems help researchers outline, draft, review, and publish evidence-backed writing.

Maya ChenMay 18, 2026CC BY 4.0
Public research article

Research question

How can AI-assisted research notes show enough process evidence for readers to trust the final article?

Findings

Transparent research writing needs more than a generic disclosure paragraph. The useful disclosure layer connects each AI action to a concrete purpose, timestamp, model, and article version.

VeriScholar treats AI work as reviewable provenance. Outline generation, draft assistance, and peer review each become durable usage records. Citation status and version history then give readers a compact way to inspect what is finished, what is still provisional, and who reviewed it.

Proposed workflow

  1. Authors draft in Markdown.
  2. AI actions are logged as structured events.
  3. Citations move through missing, needs review, and verified states.
  4. Human reviewers promote articles from AI reviewed to human reviewed.
  5. Published articles expose disclosure, license, version, and citation status.

Limitations

AI review is not a substitute for expert review. It is a consistency and risk-detection layer that should make the human review process easier to audit.