Habituation at the Gate: Rising Approval and Declining Scrutiny in Human Review of AI Agent Code

2026-06-21

AI Summary

This study investigates whether human reviewers progressively lower their standards when repeatedly evaluating AI-generated code over time. Leveraging the AIDev dataset, we conduct a longitudinal analysis of 11,429 review actions by 400 repeat reviewers over seven months, employing paired-sample Wilcoxon signed-rank tests, quantile grouping, and multivariate controls. We report the first evidence of experience-driven reflexive habituation among reviewers: approval rates increased from 30.1% to 36.8% (+14.5 percentage points cumulatively), accompanied by a 22% reduction in comment volume and a 3.5-fold increase in review latency, collectively indicating significantly diminished review effort. This effect persists after accounting for temporal trends and task difficulty, revealing a novel mechanism of human oversight decay in AI-augmented software development.

Abstract

As AI coding agents (e.g., GitHub Copilot, Devin, OpenAI Codex, Cursor) submit pull requests to open-source repositories at scale, a key question arises: do human reviewers gradually lower their scrutiny for AI-generated code over time? We conduct a longitudinal within-reviewer analysis using the AIDev dataset, studying 400 repeat reviewers who collectively submitted 11,429 reviews over a seven-month observation period. Comparing each reviewer's early and late review episodes, we observe a population-level shift in approval rate from 30.1% to 36.8% (Wilcoxon signed-rank p < 10^{-6} on paired shifts). Pooled by within-reviewer experience decile, the cumulative gap reaches +14.5 pp from first to tenth decile. This shift is experience-driven (persists after controlling for calendar time), agent-specific (human PR approval rates decline over the same period), and not explained by PR difficulty (median PR size is flat). However, review latency increases rather than decreases (+3.5x), while inline comment volume decreases (-22%, p=0.0014), suggesting reviewers spend more time in queue but less time actively inspecting code. The combination of rising approval, declining comment effort, and increasing queue time is most consistent with reflexive habituation under growing workload rather than rational trust calibration alone.

Problem

• AI coding agents
• code review
• habituation
• approval rate
• scrutiny

Innovation

• habituation
• AI code review
• longitudinal analysis
• human-AI collaboration
• trust calibration