The online systems that make recommendations to us often rely on their digital footprint — our clicks, views, purchases, and other digital footprints — to infer our preferences. But this means that human biases are baked into the algorithms. To build algorithms that more effectively predict users’ …