AI-Assisted Peer Review¶
AI-assisted peer review uses LLMs and specialized tools to simulate reviewer feedback, identify issues in manuscripts, and support the editorial process — complementing but not replacing human peer review.
Context & Background¶
The peer review system faces growing strain from increasing submission volumes and reviewer fatigue. AI can help by:
- Pre-submission review: Authors use AI to get feedback before submitting
- Desk review support: Editors use AI to quickly assess scope and quality
- Reviewer assistance: AI identifies potential issues for human reviewers to investigate
- Referee report generation: Tools like Refine.ink that produce structured manuscript feedback
Practical Implications¶
- Use AI feedback before submission: Get AI critique to improve your manuscript before human reviewers see it
- Don't submit AI reviews as your own: If you're a reviewer, AI can assist but the review should reflect your judgment
- Be aware of sycophancy: AI reviewers tend to be too positive — explicitly ask for criticism
- Complement with human feedback: AI review is a supplement to, not substitute for, colleague feedback
Key Sources¶
- The Train Has Left the Station: Agentic AI and the Future of Social Science Research
- Prompt, Plan, Review, Revise
- Project APE: Can Policy Evaluation Be Automated?
- My Claude Code Setup - Academic Workflow Template
- How Do Scientists Use Claude Code?