Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Beyond Personhood: Agency, Accountability, and the Limits of Anthropomorphic Ethical Analysis

Jessica Dai

TL;DR

The paper interrogates two dominant theories of agency—mechanistic and volitional—and their implications for AI ethics and accountability. It argues that the mechanistic view treats AI as potential moral agents but fails to ground accountability, while the volitional view rejects AI agency as intrinsic, reframing AI as the outcome of political processes. The authors propose two practical alternatives: applying agency analysis to specific applications for clearer correctness standards, and treating AI as an artifact shaped by human governance, thereby foregrounding legitimacy, process, and democratic participation. This framing aims to shift accountability from the supposed moral character of AI to the political and procedural contexts that produce and regulate AI systems, with implications for design, governance, and public trust.

Abstract

What is agency, and why does it matter? In this work, we draw from the political science and philosophy literature and give two competing visions of what it means to be an (ethical) agent. The first view, which we term mechanistic, is commonly--and implicitly--assumed in AI research, yet it is a fundamentally limited means to understand the ethical characteristics of AI. Under the second view, which we term volitional, AI can no longer be considered an ethical agent. We discuss the implications of each of these views for two critical questions: first, what the ideal system ought to look like, and second, how accountability may be achieved. In light of this discussion, we ultimately argue that, in the context of ethically-significant behavior, AI should be viewed not as an agent but as the outcome of political processes.

Beyond Personhood: Agency, Accountability, and the Limits of Anthropomorphic Ethical Analysis

TL;DR

The paper interrogates two dominant theories of agency—mechanistic and volitional—and their implications for AI ethics and accountability. It argues that the mechanistic view treats AI as potential moral agents but fails to ground accountability, while the volitional view rejects AI agency as intrinsic, reframing AI as the outcome of political processes. The authors propose two practical alternatives: applying agency analysis to specific applications for clearer correctness standards, and treating AI as an artifact shaped by human governance, thereby foregrounding legitimacy, process, and democratic participation. This framing aims to shift accountability from the supposed moral character of AI to the political and procedural contexts that produce and regulate AI systems, with implications for design, governance, and public trust.

Abstract

What is agency, and why does it matter? In this work, we draw from the political science and philosophy literature and give two competing visions of what it means to be an (ethical) agent. The first view, which we term mechanistic, is commonly--and implicitly--assumed in AI research, yet it is a fundamentally limited means to understand the ethical characteristics of AI. Under the second view, which we term volitional, AI can no longer be considered an ethical agent. We discuss the implications of each of these views for two critical questions: first, what the ideal system ought to look like, and second, how accountability may be achieved. In light of this discussion, we ultimately argue that, in the context of ethically-significant behavior, AI should be viewed not as an agent but as the outcome of political processes.
Paper Structure (25 sections, 1 figure)

This paper contains 25 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Why agency?
  3. What do we want from a framework on agency?
  4. Mechanistic Agency: A Common View in AI Practice
  5. Defining mechanistic agency
  6. The definition, in theory
  7. The implicit view of AI practice
  8. LLMs as mechanistic (moral) agents?
  9. Mechanistic agency: implications for ethics and responsibility
  10. Ethics as an epistemological question; AI as a simulator of an ideal person
  11. Can AI as a mechanistic agent be held responsible?
  12. Volitional Agency: an Alternative Approach
  13. Defining the volitional view of agency
  14. The definition, in theory
  15. An incompatible view for AI practice
  16. ...and 10 more sections

Figures (1)

  • Figure 1: A summary of the core arguments made in this paper. There are two core views of human agency; the mechanistic view is commonly assumed by AI research, yet leads mostly to a dead-end for establishing accountability. (Specifically, AI may be a mechanistic agent, but cannot be considered a moral agent even under the mechanistic view.) On the other hand, the volitional view of agency disqualifies AI itself from being considered an agent, but it allows us to view AI instead as the outcome of a political process, and therefore engage in accountability in terms of legitimacy and correctness.