We think we’re making active choices. But, often, we’re reinforcing unconscious biases - toward what’s familiar and frictionless. The algorithm isn’t evil. It’s obedient. It responds to what we train it on.