The Trust Budget: AI has to earn trust before there is any to spend
Before an AI does anything for you, it has to earn your trust.
I published an article, The Trust Budget, on Humorphism.com. Humorphisn is the design philosophy of my organization for the AI era.
Some background first. In earlier interface design there was a principle called skeuomorphism: software was made to resemble physical objects so people would know how to use it. A digital calculator looked like a real calculator, a notes app looked like a paper notepad.
Humorphism is a design philosophy that takes the next step. Instead of imitating objects, it models people, specifically the way good human partners work together, so that an interface can build the same kind of working rapport with us.
Good collaborators share a set of habits, perhaps ten of them: confirming before they act, speaking up when they are unsure, stepping in when the other person is stuck. Humorphism is about designing those habits into the way AI and people work together. My article is about what comes before that, because before an AI does anything for you, it has to earn your trust.
What I found is that the deciding factor for trust building is not how large or complex a task is. It is risk: the cost of a mistake and how hard it is to undo. We hand over small, reversible work quickly, and we hold back on anything with real consequences until trust is established.
Read the full article on Humorphism:
The Trust Budget - https://humorphism.com/news/the-trust-budget
A Chinese version is here.