OpenAI is transforming its Safety and Security Committee into an independent “Board oversight committee” that can hold off on model launches if there are safety issues.
