AI Code Review System Uses Philosophical Dispositions to Improve Bug Detection
An innovative AI-driven code review framework influences reviewer conduct through various philosophical perspectives, such as Pyrrhonian Skepticism, Navya-Nyaya reasoning, Diogenes' Cynicism, and Confucian ethics. Each perspective highlights distinct issue types, characterized apophatically (by what it does not address) and featuring a self-regulating failure mechanism (hamartia). The system operates sequentially according to role protocols. Tested on 50 combined pull requests from 7 repositories across 5 programming languages (Python, Go, C++, Java, Terraform), involving 5 organizations (2 enterprise, 3 open-source), and spanning 2 time periods (pre-AI 2020, post-AI 2024–2026), the disposition framework achieves a convergence rate of 46%.
Key facts
- System uses philosophical dispositions: Pyrrhonist Skepticism, Navya-Nyaya logic, Diogenes' Cynicism, Confucian relational ethics
- Each disposition is defined apophatically (by what it refuses to do)
- Each disposition has a self-monitoring failure mode called hamartia
- Dispositions orchestrated in sequence by role protocols
- Evaluated on 50 merged pull requests across 7 repositories
- Covers 5 programming languages: Python, Go, C++, Java, Terraform
- Covers 5 organizations: 2 enterprise, 3 open-source
- Covers 2 temporal eras: pre-AI 2020, post-AI 2024–2026
- System achieves 46% convergence
Entities
—