OpenAI Adopts Preparedness Framework for AI Safety

OpenAI recently published a beta version of their Preparedness Framework for mitigating AI risks. The framework lists four risk categories and definitions of risk levels for each, as well as defining OpenAI's safety governance procedures.

ai ai risks anthony beta definitions framework generative-ai governance large language models lists ml & data engineering openai risk risks safety