Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety

Authors: Korbak, Tomek, Balesni, Mikita, Barnes, Elizabeth, Bengio, Yoshua, Benton, Joe, Bloom, Joseph, Chen, Mark, Cooney, Alan, Dafoe, Allan, Dragan, Anca, Emmons, Scott, Evans, Owain, Farhi, David, Greenblatt, Ryan, Hendrycks, Dan, others

Publication: arXiv (2025)

DOI: 10.48550/arXiv.2507.11473

arXiv: 2507.11473

URL: https://doi.org/10.48550/arXiv.2507.11473

Philosophically-motivated questions

Analysis by Charles Rathkopf Last updated: June 2026

[Questions to be written]

Abstract

AI systems that “think” in human language offer a unique opportunity for AI safety: we can monitor their chains of thought (CoT) for the intent to misbehave. Like all other known AI oversight methods, CoT monitoring is imperfect and allows some misbehavior to go unnoticed. Nevertheless, it shows promise and we recommend further research into CoT monitorability and investment in CoT monitoring alongside existing safety methods. Because CoT monitorability may be fragile, we recommend that frontier model developers consider the impact of development decisions on CoT monitorability.

Citation for this analysis

Charles Rathkopf, “Philosophical Questions in Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety,” AI Deception Papers, June 2026, https://doi.org/10.48550/arXiv.2507.11473