Researchers have proposed a three-dimensional typology for understanding the concept of agency in advanced artificial intelligence (AI) systems. The framework, developed by combining insights from philosophy, ethics, legal theory, and sociology, considers three aspects of agency: its nature (moral or legal), mode (individual or collective), and locus (human or non-human). This typology creates a nuanced understanding of the different forms of agency exhibited by AI systems, which is particularly relevant in situations where AI actions are difficult to attribute to human actors. The authors argue that this framework can help clarify the distinction between moral and legal agency in advanced AI systems. --- Why it matters: This work matters because it provides a more precise understanding of how advanced AI systems interact with humans and their environment, which is crucial for developing responsible AI systems and avoiding potential conflicts. Source: https://arxiv.org/abs/2608.20041