Defining fairness in AI requires policy decisions only leadership can make

Evidence

The leadership team will be responsible for accomplishing a number of important objectives. First, it must develop AI standards, processes, and policies. This includes addressing data privacy and protection concerns, of course, but it also involves addressing theoretical queries. Above, we discussed that there are many definitions of fairness. We also discussed that, within fairness discussions, there are many questions (such as which aspects of criminal histories to include) that require policy answers. While these thorny questions may be flagged by the data analytics team, it is the leadership team that must answer them. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 1190