This paper aims to introduce a conceptual and diagnostic model. This model evaluates how ethical responsibility is distributed in hybrid human–AI systems operating in high-stakes domains.
This study develops a multi-dimensional framework of responsibility and proposes the ethical responsibility distribution evaluation model (ERDEM). It is drawn on science and technology studies (STS), postphenomenology and ethics of care. The framework outlines three responsibility dimensions: agency attribution, responsibility types and interaction modes. ERDEM evaluates ethical design based on five criteria – transparency, accountability, fairness, role clarity and human control.
This paper identifies four system configurations that shape responsibility flows in hybrid arrangements. It demonstrates the practical value of ERDEM by illustrative cases. It shows how system design and institutional settings influence moral clarity and blame attribution.
This study shifts attention from individual blame to systemic responsibility. It contributes to emerging theories in AI governance. These include design ethics and relational agency.
ERDEM gives system designers and regulators a structured way to assess ethical robustness. It turns abstract principles into practical criteria. The framework also guides the development of governance practices that remain sensitive to responsibility. It provides a structured method to assess ethical robustness in AI-supported decisions. It also supports the development of responsibility-sensitive AI governance practices.
This study advances the debate on AI responsibility. The focus shifts from individual culpability to distributed responsibility across socio-technical systems. Such responsibility is designable and embedded within socio-technical systems. It contributes to a novel conceptual model and actionable tool for ethics-by-design in hybrid decision-making contexts.
