For a long-time, “noses in, fingers out” told boards to stay informed and stay out of operations. It worked because the people making consequential decisions were human - visible, accountable, replaceable.
But now that assumption is gone. The most consequential actors inside many organizations are now AI agents that act autonomously, before any human reviews the decision. And on June 2, a presidential Executive Order made the organization that deploys them legally accountable.
You cannot govern an agent you cannot name. You cannot constrain authority you have not mapped. And you cannot defend a decision you never documented.
Chris and I sit with this shift for about 25 minutes.
What the episode covers:
- Why a chatbot and an AI agent are completely different governance problems.
- What Section 4 of the June 2 Executive Order actually changed - and why civil exposure became criminal.
- The deployer layer - why most boards sit exactly where the most underappreciated liability lives.
- The four things you must now be able to prove, and the single best question to ask management at the next board meeting.
Let’s Get Elemental!
