Just published an essay about the AI governance gap. Take a read and comments your views:

https://elieljoseph.substack.com/p/the-ai-governance-gap?r=7...

This is exactly why we need trust infrastructure at the protocol level. Agent-to-agent identity verification is still unsolved - most deployments just assume the other agent is who it says it is.