Governance in the Age of AI Autonomy

Governance in the Age of AI Autonomy

The directors in this session won't be asked whether AI matters — they already know it does. What they will be asked is harder: whether the governance instincts that made them effective board leaders are now the very instincts creating their organization's blindspot. Drawing on NACD's own survey data — where 62% of boards are discussing AI but only 6% are measuring it — this session confronts a fundamental accountability gap hiding in plain sight.

From Moderna's 3,000 specialized agents to JPMorgan's direct-to-CEO AI reporting structure to Duke Energy's elimination of entire crisis workflows, the evidence is unambiguous: the companies building durable advantage aren't deploying more technology. They're redesigning their organizations — and their governance — around a world where learning happens beforedoing, not after.

What that requires of boards is specific: governance that moves upstream — into what systems are being trained on, who owns the artifacts of that learning, and how competing principles get resolved before an agent ever touches a customer. The question isn't whether boards understand AI. It's whether they've acknowledged that quarterly oversight cannot govern millisecond decisions — and when they started doing something about it.