AI Security & Governance · Session 7
The trade-off test
You're about to approve an AI system for your organization. Across 10 quick choices, you'll pick the version you'd actually launch. There are no right or wrong answers-only your real preferences. It takes about 2 minutes , and you'll receive a personalized insight into how you make AI decisions.
How it works: Each choice compares two versions of the same AI system, varying in human oversight,
monitoring, compliance, speed, and cost. Trust your instincts and choose the one you'd ship.
Choice 1 of 10
Which AI system would you ship?
Same system, built two ways. Pick the one you'd put your name on.
Option A
Option B
Tap a card - the next choice loads automatically.
Your result
—
What actually drove your choices
Longer bar = the attribute that swung your decisions most; the tag shows which way you leaned. An attribute can dominate your choices and still be one you keep trading away - that's revealed preference: what you do, not what you'd say.
The trade-off test · built on the Trust Perimeter model.
Stated preference is the poll. Revealed preference is this.
Stated preference is the poll. Revealed preference is this.