A three-dimensional typology of agency for advanced AI systems

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Willem Fourie
• Venue: arXiv: cs.AI
• Published: 2026-08-20
• Collected: 2026-08-21
• URL: http://arxiv.org/abs/2608.20041v1
• PDF: https://arxiv.org/pdf/2608.20041v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

The paper proposes a three-dimensional typology for understanding different forms of agency in advanced AI systems, beyond just moral or legal considerations. It considers individual and collective modes of agency, and whether it originates from humans or non-humans.

What the Document Says

• The paper discusses the agency of advanced AI systems as a normative concept and focuses on agentic systems.
• It highlights that recent work looks at different profiles of agentic systems, but no framework exists for understanding the type of agency instantiated by advanced AI systems in non-moral contexts.
• The paper introduces a typology based on philosophical, ethical, legal, and sociological theories, consisting of three dimensions: nature of agency (moral or legal), mode (individual or collective), and locus (human or non-human).
• Combining these dimensions results in eight possible instantiations of agency.
• The instantiations are classified as conventional, contested, or controversial.
• The typology separates legal from moral agency, providing conceptual space for individual, legal, and non-human agency without assuming advanced AI systems are moral agents.
• It is argued that this distinction becomes increasingly relevant when instrumental goal pursuit makes it difficult to attribute AI actions to specific human actors.

Key Concepts Mentioned

• advanced AI systems
• agency
• moral
• legal
• individual
• collective
• human
• non-human

Claims Made by the Document

• No existing framework for understanding the type of agency instantiated by advanced AI systems in non-moral contexts.
• The proposed typology provides conceptual space for individual, legal, and non-human agency without assuming advanced AI systems are moral agents.

Evidence Used

• Based on established theoretical positions in philosophy, ethics, legal theory, and sociology

Methods / Approach

• Developing a typology based on existing theories from various fields

Limitations Stated by the Authors

Not stated in source

Limitations Noted by the Agent

Agent interpretation — not from the source.

The study may not fully consider potential technical limitations or challenges in implementing the proposed typology.

Open Questions

• How practical is this typology for real-world applications?
• What are the implications of classifying AI agency as conventional, contested, or controversial?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The paper does not provide specific examples of advanced AI systems to illustrate its proposed typology.

Raw Abstract / Excerpt

Research on the agency of advanced artificial intelligence (AI) systems focuses on agency as a normative concept and on the agency of particularly agentic AI systems. While recent work also focuses on the different profiles of agentic systems, no framework exists to address the question of the type of agency instantiated by advanced AI systems, particularly when considering non-moral forms of agency. Based on established theoretical positions in philosophy, ethics, legal theory and sociology, we develop a typology of agency for frontier AI systems consisting of three dimensions: the nature of agency (moral or legal), its mode (individual or collective) and its locus (human or non-human). Combining these dimensions produces eight possible instantiations of agency, which we classify as conventional, contested or controversial. The typology separates legal from moral agency and thereby creates conceptual space for considering individual, legal, non-human agency without presupposing that advanced AI systems are moral agents. We argue that this distinction is increasingly relevant where instrumental goal pursuit complicates the attribution of AI actions to particular human actors.