For journal editors
PaperPrep is a free AI research methods mentor built by Eric Green at Duke University. Researchers talk through their study with Paige, who asks questions, explains and compares methods, and reads drafts against reporting guidelines. Paige is designed to give guidance and feedback rather than write manuscript text.
AI use disclosure audits
When an author uses PaperPrep while preparing a manuscript, they can share an AI use disclosure: a short statement for the manuscript, with a link to an audit trail that PaperPrep generates from what it recorded.
See a sample audit trail, built for an invented study.
- What the audit trail shows
- The kinds of help the author used in PaperPrep, such as drafts read against a reporting guideline, calculations run in code and literature searches, and the AI models that answered. If the author chooses, it also shows dates of use, counts, and excerpts of their conversation with Paige.
- What the author controls
- Whether to share it, which optional parts appear, and whether the link stays on. Authors cannot edit its wording, counts, dates or quoted text. Each published version is numbered and kept unchanged, so you can see the version the author cited.
- What it does not establish
- It covers only activity within PaperPrep. It does not show whether the author used other AI tools, and it does not certify the author's conduct, the manuscript's originality, or whether it meets your journal's policies.
Pilots for submission and peer review
We would like to develop these with a small number of journals:
- Disclosure records at submission
- A place in your submission system for the author's PaperPrep link, showing the version they cited.
- Reporting checks for editors
- A reading of the submitted manuscript against the reporting guideline for its design, item by item, quoting the lines each finding relies on, to help with technical screening.
- Number checks
- Whether the figures in the abstract, text and tables agree, computed in code.
- Preparation before submission
- A PaperPrep starting point built around your journal's reporting requirements, so authors arrive better prepared.
These are features to develop and evaluate together. They are not available yet. If your journal is interested in a pilot, write to support@paperprep.org.