Respecting cultural diversity in ethics applied to AI : a new approach for a multicultural governance

Abstract

Artificial intelligence seems to be part of our everyday lives. For some it represents the promise of a better world and many improvements that would be beneficial for humanity. For others, AI is seen as threat, if not an existential threat that needs to be controlled strictly. Whatever the stance, the need to regulate AI is now widely recognized. Short of legal instruments offering a specific framework for the development and use of AI, ethics has been summoned to set standards and establish guardrails. Yet, the number of documents pertaining to ethical standards for AI has increased exponentially to reach a point where it is difficult to know how to use them efficiently. These documents have mostly been issued to promote vested interests, and the setting of a universal code of AI ethics has been seen as a solution for AI global governance. If a global governance system is required to avoid negative outcomes of AI, it appears that the idea of a universal code of ethics denies the diversity of ethical standpoints based on the diversity of philosophical cultures the world is made of. Instead of offering a legitimate and efficient tool, such a solution could lead to cultural tensions between leading actors in the field of AI as it is the case between China and the United States. To avoid conflicting situations stemming from the denial of cultural diversity, it is more than ever necessary to put aside the idea of a universal code of AI