Critical Reflections on the Ethical Regulation of AI: Challenges with Existing Frameworks and Alternative Regulation Approaches

Abstract

This paper synthesizes three fundamental challenges within existing frameworks for the ethical regulation of artificial intelligence (AI) identified in recently published literature. These challenges have recently been raised by ethicists, computer scientists, and policymakers. More specifically, existing frameworks are encountering challenges such as the difficulties that come along with defining an AI, adapting to the needs of the public in participating in democratic governance, and considering the environmental impacts of these systems. After critically reviewing the three challenges, this paper proposes alternative regulatory approaches to address each of these regulatory challenges. To address the definition challenge, we propose a multidimensional approach to defining AI. We argue that a relational approach can be helpful for making visible the cultural context in which AI systems are embedded. Finally, we suggest that involving a systematic environmental studies approach in assessing AI is conducive to the development of not only more environmentally friendly AI systems but also AI technologies that can be used to address challenging global environmental issues.