The algorithms used in machine learning and other forms of artificial intelligence are remarkable in the speed at which they can analyze vast quantities of data to identify patterns and detect disease. But because they capture historical trends, those same algorithms can mirror human bias and perpetuate the disparities that already plague health care. As the use of AI expands, questions need to be answered about how it should be regulated and what happens when algorithms are designed to perform in unethical ways. How do we ensure fairness in machine learning to advance health equity? Who decides whether to optimize algorithms for profit, outcomes, or quality?