QCAIJ AI-Assisted Article Review Protocol
QCAIJ-ARP · Version 1.0 · AI-Assisted Review Index (AIRI)
The protocol combines an AI-assisted scholarly assessment with a transparent weighted index. The index is a pre-review signal for editorial triage and revision planning; it is not a peer-review verdict.
AI-Assisted Review Index
Ten dimensions are scored from 0–10 and weighted to produce one normalized 0–100 index.
| Code | Dimension | Weight |
|---|---|---|
| RQ | Research question and contribution | 15% |
| NOV | Novelty and significance | 15% |
| METH | Methodology | 15% |
| EVID | Evidence and claims | 15% |
| REPRO | Reproducibility | 10% |
| CITE | References and scholarly context | 10% |
| ETH | Research ethics and responsible AI | 10% |
| CLAR | Structure and clarity | 5% |
| LIM | Limitations and boundary conditions | 5% |
Interpretation
- 85–100: High pre-review readiness
- 70–84.99: Moderate readiness; revision recommended
- 50–69.99: Low readiness; substantial revision
- 0–49.99: Critical pre-review concerns
AI review output
The article review page displays the AIRI score as a visual graph and provides criterion-by-criterion findings, evidence excerpts, recommendations, strengths, major concerns, missing items, model confidence, and review limitations in HTML.
API endpoint
POST ?page=api_review_protocol
Authenticated editors, reviewers, publisher administrators, and super administrators can submit title, abstract, content, and an optional research object ID. The current protocol route is retained for compatibility; AI review is available through the reviewer workflow.