AI Deception Papers

Evaluating Chain-of-Thought Monitorability

Authors: OpenAI

Publication: Preprint (2025)

URL: https://openai.com/index/evaluating-chain-of-thought-monitorability/

Philosophically-motivated questions

Analysis by Charles Rathkopf Last updated: June 2026

[Questions to be written]


Abstract

OpenAI presents empirical evaluations of chain-of-thought (CoT) monitorability, studying the extent to which reasoning models’ chains of thought faithfully reveal information relevant to detecting misbehavior, and the conditions under which CoT monitoring succeeds or fails as a safety affordance.


Citation for this analysis

Charles Rathkopf, “Philosophical Questions in Evaluating Chain-of-Thought Monitorability,” AI Deception Papers, June 2026, https://openai.com/index/evaluating-chain-of-thought-monitorability/