OpenAI says users can set action boundaries for dots

OpenAI says users stay in control of dots by setting boundaries for three cases: what a dot can do on its own, when it should ask first, and what it should never do.

Safety controls are built in. Users choose which apps to connect, and a dot runs on its own cloud computer, so connecting the user’s computer is entirely optional. OpenAI describes the approach in how safety, security, and privacy are built into dots.