The Straits TimesTech
Review finds flaws in OpenAI’s AI monitoring system
External evaluators METR and Redwood Research found several new vulnerabilities in OpenAI’s internal monitoring system during a pilot review. The review highlights flaws in the system, but no further information about the vulnerabilities was provided.
This card is an AI-assisted summary of reporting published by The Straits Times. Read the original report before relying on it for medical, legal, financial, safety, or emergency decisions.