Today we trust humans to manage critical processes.

In the future we will need to trust humans to not run those processes with unreliable tools.

This isn't just about AI. How do we trust the humans who manage critical processes to follow good security practices (keeping systems patched, etc)?