That is a good point. Perhaps as we build better and better computers they might point us in the correct direction. What I mean is that data analysis programs will start to show some correlations between pieces of data that we have previously never imagined to have any connection.Regarding biases we don't observe in anyone again systemic causation is our best shot. We can sometimes figure out something must exist like dark matter or dark energy or black holes because something is missing, not because we have observed it. Now, I don't know if any of those three things actually exist.
Scientists keep going back and forth with all of them and pointing out problems if they do exist, so maybe they don't exist but SOMETHING does some of the things that dark matter, dark energy and black holes are supposed to do, even if those three things don't actually exist.
So, for example economics used to have the idea that people are rational actors. People observed human beings and found that we don't conform to models that are us as unbiased pure rational actors. So psychologists looked at people and saw the biases and behavior that has created behavioural economics.
The book Thinking, Fast and Slow by Daniel Kahneman details this as do many, many others.
So, we can find biases that are difficult to observe because biases exert influence. The influence, like the gravity of a black hold or the expansion of some things and the tendency to stay together of other things can show that something, or maybe several things seems to do some of the things that dark matter and energy are supposed to do.
So, if biases are influential they leave a trail of breadcrumbs to follow and if they exert no influence then they might be inconsequential and leave no trail.