OpenAI partnered with the Frontier Model Forum and member firms on a shared methodology to identify when bypassing an AI system’s safeguards could enable serious cyber harm and to prioritize stronger protections.
Need help?
Contact usOpenAI partnered with the Frontier Model Forum and member firms on a shared methodology to identify when bypassing an AI system’s safeguards could enable serious cyber harm and to prioritize stronger protections.