Aren't algorithms also biased in theory since they are created by people?
In machine learning it might be easier to unintentionally create biased algorithms. On the other hand there are domains where I would trust algorithms to be less biased than people.