Robbie, Klara, and Ethan: Replicability and the Moral Status of AI

What moral consideration would we owe future AI that had human-level intelligence? I address this question by drawing on philosophy and science fiction. The works of science fiction invite us to form intuitions about how we should treat AI, that can be tested against rational criteria identified by philosophers, but also help us evaluate those criteria. I draw on Mary Anne Warren’s multi-criterial framework that focuses on sentience, personhood status, and relationship value—but in addition, I propose a criterion of “non-replicability”: “bottom-up programmed” entities that are unique due to their distinct, irreplicable experiences are owed special moral consideration. I apply these rational criteria to Robbie from Asimov’s I, Robot, Klara from Kazuo Ishiguro’s Klara and the Sun, and Ethan from the science fiction television series Extant. I argue that intuitively we give greater moral consideration to Ethan and Klara than Robbie and that this is supported by rational criteria including that of non-replicability. I also address who is owed greater moral consideration between either Klara and Ethan, and a human being, if Klara or Ethan are replicable after all.