The rapid adoption of Generative AI has pushed institutions into two unhelpful extremes: blanket bans driven by fear, or mindless automation that forfeits human responsibility. Neither approach addresses the core risk of the AI era—the proliferation of unverified assertions, hallucinated citations, and work that no human stands behind.
The Doctrine of the Examiner is a manifesto and a practical framework for academic and professional work in the age of AI. Rather than judging a piece of work by its origin (whether composed by hand or machine), it establishes a strict rule: work must be judged by whether it can be examined and verified.
This document outlines:
- The Fallacy of AI Detectors: Why relying on probabilistic AI scanners fails to safeguard academic integrity and creates false accusations.
- The Taming of AI: How to transform AI from a source of risk into a reliable tool by enforcing a strict verification gate—binding every claim to primary sources and ensuring figures are recomputable.
- The Role of the Examiner: Establishing the principle that machines may propose, but named human authors must decide, verify, and bear full accountability for the output (Opus iudicat, non genus).
The full text and its primary source appendix are available at Zenodo: https://zenodo.org/records/21607195
I welcome feedback and discussion from the community on how we can implement these verification standards to preserve epistemic integrity in research.
