Redwood proposes transparency rules as architectures risk removing chain-of-thought monitorability
RyanGreenblatt · x · 2026-09-11
Redwood Research published a proposal addressing a growing safety concern, backed by Ryan Greenblatt and others:
- Core worry: architectures that push reasoning into opaque activations ("neuralese") or drop chain-of-thought entirely would remove the main handle for monitoring AI reasoning.
- Concrete case: Greenblatt argues that based on limited public evidence, Google's Astra was a concerning step in this direction, yet public information is insufficient for an informed debate on the monitorability-vs-performance trade-off.
- The proposal: companies should disclose no-CoT reasoning abilities, other monitorability evidence, and publish policies for preserving monitorability.
A rare, concrete transparency framework aimed at a specific product architecture.
Related event: Redwood Research Proposes Disclosure Rules to Protect CoT Monitorability(5 posts)→
More from Safety
- PHP2Shell exploits confirmed in the wild, attackers may use agentic AI for post-exploitation — AravSrinivas · 2026-10-27
- php2shell exploit benchmark: only 2 of 5 frontier models can port the bug — xeophon · 2026-10-27
- Interactive 32-slide AI security & privacy lecture for lawyers, now public — chrisrohlf · 2026-10-27
- Two CVSS 9.9 Wasmtime flaws allow sandbox escape, one ARM64-only since 2020 — basedjensen · 2026-10-27
- Free prepare.aiendgame.com site guides users through locking down accounts amid AI cyberstorm — harris_edouard · 2026-10-27
- Cisco routers trust eMudhra CAs whose private keys were numbers under 256 — murat_kocaoglu_ · 2026-10-27