Collective responsibility under AI network: Criticism of electronic person based on neocybernetics theory

Abstract

The aim of this paper is to consider how to position the attribution of ethical responsibility in situations where AI and robots are widely used in society and in situations where cooperation exists among multiple AIs and between AIs and other systems. Based on the theory of neocybernetics, we express concern about the electronic persons' proposal introduced in the European Parliament and support the idea of collective responsibility under the AI network environment. The electronic persons' proposal involves according a position as a personality to what is not an autopoietic system. It does not match the actual situation and causes ethical problems. Even if we do not establish an electronic person institutionally, we can compensate for it by establishing systems of collective responsibility. In view of recent trends in computer technology, we assume that responsibility cannot be attributed to specific people or organizations. In such a case, establishing collective responsibility is required to compensate the victims and not give rise to a situation wherein people do not favor developing or using AI. However, under the AI network situation, there is lack of collective responsibility. If the defects of a specific person or organization are clear, responsibility should be attributed to the concerned person or organization. This has been the custom after the modern era and one that is difficult to change. Besides, to prevent deliberate crimes, negligence, and decline in the responsibility of developers/users and to remain motivated to improve this technology, it is necessary to impose ethical responsibility on wrongdoers, whether they are people or organizations.