Fair enough. I'm definitely just drawing from personal anecdotes and logic for what I think society would do. Just like everyone. Even though studies are done on societies, I personally think it's hard to factor in every important variable. And since most studies these days just skew their data to try and prove a point, it's hard to trust in those either.
I'd be fascinated to see such a model of society, it'd be incredibly helpful in economics, finance, social policy, education, fitness, law, etc, etc.
I suspect if you could make a machine that could predict human behavior that accurately, you'd want to keep it to yourself, while you became unfathomably rich and powerful.
That is why I pointed out colleges as a case study. The thing is, colleges weren't always that way. Sure students have always played pranks or messed around (I think of John Quincy Adams at Harvard getting caught streaking on the lawn drunk), but it's clear that sort of thing was uncommon. Compare that to today, when the only thing that changed is the way society treats college. I'm not sure how we can predict and change things in people's lives enough to make society as a whole take college seriously.
Compare how simple that is to the problem BI is trying to solve: some people can't find work. It boggles the mind contemplating how big an issue that is, and how much we'd have to change people for it to work.