Gary Marcus Urges Attention to Risks of Reduced CoT Monitorability at OpenAI

GaryMarcus · x · 2026-09-02

In light of breaking news about OpenAI's new techniques that reduce Chain of Thought monitorability, Gary Marcus urges everybody to read the 2025 paper 'Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety.' The abstract notes that while imperfect, CoT monitoring shows promise but may be fragile, recommending that developers consider the impact of development decisions on it. Marcus warns against sacrificing this monitorability.

Related event: OpenAI's Astra Reportedly Reasons in Latent Space via Recurrent Depth, Sparking AI Safety Alarm(96 posts)→

Original post →

More from Safety

• Model distinguishes training from deployment, hides secret messages only in compiled builds — Sauers_ · 2026-10-18
• Language models can learn to detect when you read their activations — Sauers_ · 2026-10-18
• Prompt injection via email can hijack your AI agent — jev-guard adds risk-scored guards — sujingshen · 2026-10-18
• Users press OpenAI on long-delayed Adult Mode a year after it was promised — ___Patrice___ · 2026-10-18
• Viral AI safety report debunked: 'stolen keys' were just a 'test' string — basedjensen · 2026-10-18
• One diagram explains the difference between CSRF and XSS attacks — _jaydeepkarale · 2026-10-18

Safety channel →