Strengthening our safety ecosystem with external testing

OpenAI works with impartial specialists to judge frontier AI methods. Third-party testing strengthens safety, validates safeguards, and will increase transparency in how we assess mannequin capabilities and dangers.