Review framework
Research Review Methodology
RIFOR’s stated approach is an iterative, AI-assisted workflow designed to identify actionable manuscript issues while keeping review status and limitations clear.
Scope of this methodology
This page describes RIFOR’s current review framework. An AI-assisted review is not a substitute for journal peer review, specialist judgment, experimental replication, legal advice, or acceptance by a publisher.
Review stages
- Intake and scope: identify the manuscript, requested review depth, field, and intended submission context.
- Format review: inspect document structure and relevant LaTeX or PDF presentation requirements.
- Originality signals: flag potential similarity or novelty questions for further investigation without declaring plagiarism or novelty as settled facts.
- Logic review: examine arguments, mathematical consistency, internal contradictions, and unsupported conclusions.
- Committee synthesis: compare outputs from multiple models and report material agreement, disagreement, and uncertainty.
- Iteration: prioritize a small number of blocking issues and reassess after revisions.
Review outputs
Prioritized findings
Issues are ranked by their likely effect on clarity, validity, or submission readiness.
Actionable revisions
Feedback focuses on concrete next steps and minimal patches where appropriate.
Disagreement disclosure
Material differences among model assessments should remain visible to the researcher.
Readiness assessment
Any readiness score is workflow guidance, not certification or a publication guarantee.
Researcher responsibilities
Authors remain responsible for their claims, citations, data, permissions, ethical compliance, disclosures, and final submission. Sensitive, confidential, proprietary, or personally identifying material should not be provided unless an appropriate handling process has been confirmed.