I also wanted to pass along this explanation from our team manager, Jack Clark:
OpenAI’s policy team looks for candidates that display an ‘idiosyncratic specialism’ along with verifiable interest and intuitions regarding technical aspects of AI technology; members of OpenAI’s team currently have specialisms ranging from long-term TAI-oriented ethics, to geopolitics of compute, to issues of representation in generative models, to ‘red teaming’ technical systems from a security perspective, and so on. OpenAI hires people with a mixture of qualifications, and is equally happy hiring someone with no degrees and verifiable industry experience, as well as someone with a PHD. At OpenAI, technical familiarity is a prerequisite for successful policy work, as our policy team does a lot of work that involves embedding alongside technical teams on projects (see: our work throughout 2019 on GPT2).
I also wanted to pass along this explanation from our team manager, Jack Clark: