"Power, Process, and Automated Decision-Making" by Ari Ezra Waldman is a good discussion on how biases manifest in algorithmic decision-making. I agree that an

"Power, Process, and Automated Decision-Making" by Ari Ezra Waldman is a good discussion on how biases manifest in algorithmic decision-making. I agree that an algorithm would be agnostic to the data, but that biases inherent in the data would be cemented, not overcome, by agnostic algorithms.