反思機器人的道德擬人主義

Reflection on the Moral Anthropomorphism of Robots

如果機器人的發展要能如科幻想像一般，在沒有人類監督下自動地工作，就必須確定機器人不會做出道德上錯誤的行為。根據行為主義式的道德主體觀，若就外顯行為來看，機器人在道德上的表現跟人類一般，機器人就可被視為道德主體。從這很自然地引伸出機器人的道德擬人主義：凡適用於人類的道德規則就適用於機器人。我反對道德擬人主義，藉由史特勞森對於人際關係與反應態度的洞見，並以家長主義行為為例，我論述由於機器人缺乏人格性，無法參與人際關係，因此在關於家長主義行為上，機器人應該比人類受到更嚴格的限制。

If robots are to function automatically, without human supervision, as depicted in sci-fi imagination, then we must ensure that robots not commit moral wrongs. According to the behaviourist conception of moral agency, if robots, assessed purely on the basis of behaviour, per-form as morally as humans, they can be considered moral agents. This naturally leads to moral anthropomorphism: the position that whatever moral standards apply to humans apply equally to robots. I argue against moral anthropomorphism. In light of P. F. Strawson’s insights into interpersonal relationships and reactive attitudes, and drawing on paternalist actions as examples, I argue that robots, being not persons, are unable to participate in interpersonal relationships, and therefore their paternalist actions towards humans ought to be less permissible than humans’.