How much should you trust AI?

Click the slide to open it full size.

From the talk

Let me ask you a question. How do you know you trust someone? You rely on them.

You let yourself be vulnerable. You stop checking every action. Now think about AI.

Some people say I never trust technology, and they miss out on real help. Others say the system said so, and they stop thinking. The goal sits in the middle: I know when and how to verify.

That is what the rest of tonight is about.

Read the slide as text

The goal is not more trust. The goal is better calibrated trust.
BLIND REJECTION
CALIBRATED TRUST: THE GOAL
BLIND RELIANCE
“I never trust technology.”
“I know when and how to verify.”
“The system said so.”
When you trust someone, you rely on them, you become vulnerable, and you stop checking. But trusting behavior does not always mean good judgment.

Tip: use the ← and → keys to move between slides.