[2E5-OS-4b] OS-4

[2E5-OS-4b-05] Ethical Considerations on the Limits of Using Commonsense Morality in AI Ethics: From the Perspectives of Metaethics and Animal Ethics

〇Masashi Takeshita¹ (1. Graduate School of Information Science and Technology, Hokkaido University)

Keywords:Speciesism, Animal Ethics, AI Ethics, Commonsense Morality

This study discusses the limitations of relying on commonsense morality in the development of moral AI from the perspectives of ethics, particularly metaethics and animal ethics. In the first half of this study, I examine the general problems of commonsense morality from a metaethical perspective. First, commonsense morality varies across cultures and individuals and is ambiguous, making it difficult to identify the appropriate commonsense morality that should be implemented in AI. Second, even if a clear, commonsense morality existed, it would likely contain errors as long as it is based on moral realism. In the second half of this study, to support the second point, I argue from the perspective of animal ethics, focusing on speciesism. Speciesism refers to the unjustified comparatively worse consideration or treatment of those who do not belong to a particular species. Based on arguments in animal ethics, I maintain that sentient nonhuman animals are subjects of moral consideration and argue that justifying speciesism is difficult. However, since many people today share speciesist beliefs, commonsense morality contains speciesism, which is a morally wrong idea. For these reasons, I conclude that there are problems with assigning a significant role to commonsense morality in AI ethics.