[schoen]

Cf. https://arxiv.org/abs/1410.8233 ("Do Artificial Reinforcement-Learning Agents Matter Morally?").