But why?: understanding explainable artificial intelligence

Opaque algorithms get to score and choose in many areas using their own inscrutable logic. To whom are said algorithms held accountable? And what is being done to ensure explainability of these algorithms?