Five Policy Uses of Algorithmic Explainability

The notion that algorithmic systems should be "explainable" is common in the many statements of consensus principles developed by governments, companies, and advocacy organizations.