Hi Here is a question for all: a) Should we treat AI agents as any other identity/device etc. b) What access control apart from authorization/authentication, do you guys feel should be added to an agent. c) Should this be integrated into a current product like IAM or a separate product make more sense.

I have been thinking about identity for agents itself

We can treat it as a machine like identity (System) or a user like identity. I am leaning towards user like identity. But users rarely (number wise) go rogue, so we have to find out how to limit the rogue-ness (intentional/un-intentional)