In accident scenarios, potential AI ethical programming includes Retributivist (harm responsible party), Selfish (protect AI's occupant), and Utilitarian (minimize overall harm). Each has implications. #AIethics #DecisionMaking