Introduce Superalignment
We have a number of ideas that have been tested and shown to be promising. We have a growing set of metrics that we can use to measure progress. And we can study these problems using today’s models.
Ilya Sutskever, cofounder and Chief Science Officer of OpenAI, has set this as his primary research focus and will lead the team along with Jan Leike. Researchers and engineers from the previous alignment team as well as other researchers across the company will be joining the team.
We are also seeking outstanding new researchers, engineers and scientists to join our effort. We believe that superintelligence alignment is a fundamental machine learning problem. Great machine learning experts, even if they are not working on alignment already, will be crucial to solving this.
Source:
https://openai.com/blog/introducing-superalignment
