Six security measures that we believe will complement today's security controls and help protect advanced AI:

→ View original post on X — @openai