Reinforcement learning and artificial agency