Is Trust Measureable?

by Nam Chu Hoai about 9 years ago

A Response to "Can Trust Systems Build a New Economy From Ruin?"

Shareable just had a great article about trust - this post won't make sense unless you read it here. The article discusses a handful of very difficult questions. There's an extended discussion of FICO in comparison to trust. We thought a trust score deserved a post in itself, since it seems to be a popular approach.

In order to create a global trust system, one has to define what trust is. A reasonable attempt to quantify something, like some try to do with a trust score, has to lay out exactly what is being quantified.

A quick look to Merriam Webster brings "assured reliance on the character, ability, strength, or truth of someone or something". One of our previous blog posts also has a lengthier discussion in attempting to define trust. In my mind, trust is (confidence in) our ability to predict future behavior. For the sharing economy, it comes down to whether or not you'll have a good experience when someone stays in your home, drives your car, or any other interaction.

When you look at trust in this context, the differences between FICO and a trust score become immediately apparent. FICO only predicts one thing: your ability to pay back debts. A credit score can be reasonably accurate based on a few numerical attributes - your past history of paying debts, your current debt load, the length of your credit history, etc. It also has the benefit of very measurable outcomes. Did you pay your debts back? How quickly?

In contrast, a trust score requires the prediction of a seemingly endless list of behaviors. Are you a good driver? Are you courteous when staying in someone's home? Do you take care of dogs well? Are you dangerous around my children? Will you do a good job putting together my IKEA furniture? It seems hard enough to gather data on one person in all of these situations, let alone predict future behavior in any of them.

Even more important than the prediction itself is the error associated with a false prediction. In credit scores, a score that's artificially high will in the worst case mean lost money for the borrower. A credit score which is too low will only lead to opportunity cost.

On the other hand, a trust score which is too high won't just jeopardize money. It potentially puts human life in danger to blindly trust someone malicious. If we as a society become dependent on a score, every false prediction threatens human safety, and irreparably damages the scoring agency itself. Erring with a score that's too low might not threaten safety, but it does shut off participation in sharing marketplaces, which is the last thing they need right now.

In short, one score can't encompass the vast number of situations, contexts, and definitions which fall under the term trust. Even if some score did manage to come reasonably close to predicting human behavior, every false prediction would have such a negative effect that the score would be unsustainable.

comments powered by Disqus