CNA·Tech

METR finds flaws in OpenAI monitoring system

Summarised by Spiced Kopi · 22 October 2026, 16:02 UTC ·8 sec listen

AI safety group METR says it found flaws in OpenAI’s internal monitoring system. The available information does not explain what the flaws are, how serious they may be, or whether OpenAI plans to change its monitoring approach.

This card is an AI-assisted summary of reporting published by CNA. Read the original report before relying on it for medical, legal, financial, safety, or emergency decisions.