Priorities and principles for effective third party assessments

Source: OpenAI

What was announced

OpenAI outlines priorities and principles for rigorous, secure, and independent third-party AI safety assessments of frontier models and safeguards.