Who should bear moral responsibility when autonomous AI makes decisions?

Details

As AI systems become more autonomous and adaptable, they challenge traditional boundaries of moral responsibility in techno-social systems. This conceptual paper explores the evolving discussion about delegating responsibilities to intelligent autonomous agents and the ethical implications of doing so.

Its method is a synthesis of recent developments in AI ethics, including distributed responsibility and ethical AI by design, from which it proposes a functionalist framework. In this perspective, moral responsibility is not an individual trait but a role within a socio-technical system, distributed among human and artificial agents.

The paper presents Basti and Vitiello's implementation as an example of “AI ethical by design.” They suggest that AI can act as artificial moral agents by learning ethical guidelines and using Deontic Higher-Order Logic to assess decisions ethically.

Motivated by the possibility that AI could operate at speeds and scales beyond human supervision, and by the ethical implications, the paper argues for “AI ethical by design.” At the same time, it acknowledges that responsibility is distributed, shared, and dynamic. It presents this functionalist approach as a practical framework for navigating the complexities of AI ethics in a rapidly evolving technological landscape.