So yes, people more or less react the same way as computer models, but that’s what makes those models so useful, they can act as a control in a way the real world can’t. And to that end we actually hope to see deviations, because that means we didn’t program every variable in to the program. And by discovering what the missing variable is we can often discover the root of societal problems.
When we discover that this deviation affects one particular group to a much greater degree than expected we begin to consider the possibility of a bias in the system itself, or systemic bias. Systemic bias can take all kinds of forms, for instance if one were to look at LGBT demographics across the country we see higher densities in certain cities like Miami and San Francisco than expected. Most of us would consider that a positive bias, but it is a bias, specifically a sexual orientation bias. Or you can look at income levels across a Fortune 500 companies at which point you’ll find a bias resulting in women making less money for performing the same job, a gender bias. The general term for this is system or systemic bias. When the bias focuses on a particular racial group then it’s racial bias, or systemic racism.
But systemic racism or racial bias isn’t the end all be all. It’s the variable that causes the system to behave in a way we’re isolating for. The next step is to determine what is causing that bias. When we look into it deeper we see noticeable trends of poor education, low income, decreased opportunites for socioeconomic advancement or even survival (or the perception of decreased opportunities), etc. Many of these are also the result of historical racial biases. But much like an alcoholic first needs to admit that they have a problem, society first needs to recognize the existence of systemic racism before we can dig deeper to discover the roots of that bias and find solutions to those underlying issues.