The Moral Development of Artificial Intelligence
Le développement moral de l’intelligence artificielle
This article examines the moral development of artificial intelligence, exploring both theoretical dimensions and practical implications of implementing ethical principles in AI systems. Starting from the fundamental distinction between two architectures – Combinatorial-state automata (CSA) and Level of Abstraction (LoA) – the study investigates how AI systems can incorporate and exhibit behaviours aligned with moral principles. Through comparative analysis of two concrete AI implementations – the „Cătălin” call centre robot and the Character.ai platform – the article demonstrates significant differences between CSA systems, which strictly pursue predefined objectives, and LoA systems, which incorporate ethical thresholds and adaptive capabilities. In the current context of accelerated artificial intelligence development, the paper argues for the fundamental necessity of integrating moral databases into AI system architecture, emphasizing the significant impact of this approach in both educational and commercial interactions. The conclusion highlights that, while we cannot yet speak of an intrinsic morality in artificial intelligence, implementing ethical principles in AI system programming represents an imperative necessity for developing human-machine interactions based on trust and moral values.