The fundamental issue is that the stranger is proving something -- the fact that they are close to your family -- by supplying knowledge -- the answer to the qu

The fundamental issue is that the stranger is proving something -- the fact that they are close to your family -- by supplying knowledge -- the answer to the questions. Because he or she is proving something by supplying knowledge, this cannot be viewed as a "zero-knowledge proof". The minimal example of what would have to change would be something along the lines of this: rather than asking specific questions, you would want the stranger to be able do something that only a family friend could do. But, in this case, it's silly to get them to do that when what you want is that knowledge -- who they know, how they know them, etc. -- so you can verify it against what you already know. The question of whether or not someone is a family friend is just not the type of question that these proofs are used for. I would throw out that example and use something like this: a lock box that anyone could close -- but only someone with knowledge of a secret code could open it. Someone could prove that they know the code by simply opening the box. It's a contrived example, yes, and could probably use some extra pizazz, but I think that's because you'll find most real-world situations simply don't work well with these sorts of proofs. Also, is somewhat relevant: https://news.ycombinator.com/item?id=14261067 https://news.ycombinator.com/item?id=14261067