Criminal liability in AI-assisted suicide: Theoretical challenges and doctrinal responses

The growing phenomenon of AI-assisted suicide poses a profound challenge to the traditional doctrines of criminal attribution. The difficulty becomes particularly acute when an AI system provides lethal information, reinforces suicidal ideation, or directly facilitates the fatal act. Because contemporary AI systems cannot themselves bear criminal liability, the analysis must focus on the natural persons and, where applicable, corporate entities involved in their design, deployment, operation, and use. Moving beyond static models of liability, this article develops a differentiated, role-based framework of criminal attribution. Under this framework, developers and platform operators may bear continuing duties relating to safety-conscious system design, proactive risk mitigation, and supervision throughout the system's lifecycle. At the same time, an individual's autonomous self-harm remains, in principle, beyond the reach of criminal punishment; users incur criminal liability only where they intentionally or negligently facilitate another person's suicide under the conditions prescribed by the applicable law. The framework further refines negligence-based imputation by distinguishing factual causation from normative attribution. Liability requires both that the fatal outcome constitute the realization of a legally impermissible risk created by the relevant breach of duty and that counterfactual compliance would have afforded the legally requisite prospect of averting that outcome. Ultimately, this differentiated framework preserves the human-centered foundations of criminal liability without treating algorithmic complexity either as a basis for attributing fictional culpability to AI systems or as a categorical barrier to the attribution of responsibility to human and corporate actors.