[White Paper] Language model safety and misuse: lessons learned (openai.com)

Describes findings and recommendations for AI developers on addressing safety risks and preventing misuse of deployed language models.