We're going to talk about five common types of algorithmic bias we should pay attention to: data that reflects existing biases, unbalanced classes in training data, data that doesn't capture the right value, data that is amplified by feedback loops, and malicious data. [10:56]