Navigating Peer Review with AI.

Pre-Submission Review

The best time to address reviewer concerns is before submission. AI can simulate the review process, identifying weaknesses that real reviewers will catch. ``` PRE-REVIEW PROMPT: Act as three peer reviewers for {journal_name} reviewing this manuscript. Each reviewer has a different focus: REVIEWER 1 (Methodology): Evaluate the study design, statistical approach, sample size, and internal validity. Identify specific methodological weaknesses and suggest improvements. REVIEWER 2 (Theory/Literature): Evaluate the theoretical framing, literature coverage, and how well the findings connect to existing knowledge. Identify missing references and theoretical gaps. REVIEWER 3 (Presentation): Evaluate clarity, structure, figure quality, and whether the paper tells a compelling story. Identify confusing sections and suggest restructuring. For each reviewer, provide: - MAJOR CONCERNS (would prevent acceptance) - MINOR CONCERNS (should be addressed but not blocking) - SPECIFIC SUGGESTIONS (actionable improvements) Be harsh. Real reviewers will be. ``` Run this prompt three times with different temperature settings (0.3, 0.7, 1.0) to get diverse critiques. The low-temperature run finds obvious issues. The high-temperature run surfaces creative objections you might not have considered.

Anticipating Discipline-Specific Critiques

Different fields have different review cultures. AI can be calibrated to your discipline: ``` DISCIPLINE-SPECIFIC PROMPT: In {field}, the most common peer review critiques are about: {paste_known_common_critiques} Review my manuscript specifically for these known patterns. For each potential critique, indicate: 1. Where in my manuscript this weakness exists 2. How severe it is (fatal flaw vs. minor quibble) 3. How to address it before submission Common critiques by field: PSYCHOLOGY: "underpowered," "no pre-registration," "WEIRD sample," "effect size not reported," "p-hacking concerns" COMPUTER SCIENCE: "no baselines compared," "single dataset," "no ablation study," "novelty unclear relative to [method]," "reproducibility concerns" BIOLOGY: "n too small," "no replication," "inappropriate controls," "overclaimed from correlation," "missing negative controls" ECONOMICS: "endogeneity," "omitted variable bias," "instrument validity," "external validity concerns," "robustness checks" ``` For each critique the AI identifies, prepare your defense before submission. Either fix the issue in the manuscript or add a paragraph in the limitations/discussion explaining why the critique is addressed or acknowledged.