The Application of the Theory of Moral Blameworthiness in Psychology to the Determination of Punishment for Crimes Arising from Artificial Intelligence Systems

Objective: This study aimed to examine how the theory of moral blameworthiness can be applied to determining criminal responsibility and punishment for crimes arising from artificial intelligence systems.

Methods and Materials: A descriptive–analytical and interdisciplinary methodology was used. The study drew on moral psychology, philosophy of responsibility, criminal law theory, AI ethics, and contemporary AI governance frameworks. Academic literature, legal instruments, policy documents, and regulatory materials were analyzed, including sources related to moral responsibility, blame attribution, AI autonomy, the responsibility gap, and risk-based AI governance.

Findings: The analysis showed that current AI systems should not be treated as fully blameworthy moral or criminal agents because they lack consciousness, moral understanding, free will, and genuine responsiveness to moral reasons. However, AI-related harm does not eliminate responsibility. Responsibility should be assessed across the human and institutional chain surrounding the AI system, including designers, developers, data providers, model trainers, deploying institutions, corporate managers, users, and regulators. The study identified knowledge, control, foreseeability, preventability, and risk creation as the main criteria for assessing blameworthiness. Traditional criminal law concepts such as mens rea, negligence, recklessness, causation, and culpability remain necessary but require reinterpretation in light of algorithmic autonomy, opacity, distributed decision-making, and organizational responsibility.

Conclusion: Punishment for AI-related crimes should be proportionate to morally relevant culpability rather than mere causal involvement. A chain-of-responsibility model can help prevent both unjust impunity and unjust over-punishment.