Title Pending 16902

If artificial intelligence (AI) programs start manifesting religious interest or behavior, how could we know whether it is genuine? This article explores the necessary conditions for robots to develop authentic religiosity from theological, evolutionary, and cognitive angles. Theology is more open to the hypothesis of religious robots than naturalistic approaches, which regard the emergence of human religion as highly contingent on the idiosyncrasies of our embodiment, needs, cognition, and evolutionary history. Drawing on Robin Dunbar’s work and educated guesses about the mental world of future AI, I argue that common human intuition about potential robot religiosity—as captured in sci-fi and pop culture—is plagued by two fallacies: (1) a bias to equating religion with only its intellectual or doctrinal layer at the expense of the embodied, shamanistic dimension, and (2) a naïve assumption that intelligent robots will develop humanlike thoughts, aspirations, and concerns. Due to its alien type of intelligence, AI will likely not engage in something so typically human as religion.