What I Look For
A practical framework for finding the gap between what an AI system is supposed to do and what it actually does when the stakes are real—and who has the least room for error.
A practical framework for finding the gap between what an AI system is supposed to do and what it actually does when the stakes are real—and who has the least room for error.
The obvious bad actor is only half the equation. The quieter failure is a system that turns help, support, or validation into an echo chamber. The harm usually lives in the arc of the conversation, not in one obviously bad turn.
For the people building, evaluating, governing, and living with the consequences of AI systems—especially the users with the least room for error.