Graduation Year
2025
Date of Submission
4-2025
Document Type
Campus Only Senior Thesis
Degree Name
Bachelor of Arts
Department
Philosophy
Reader 1
Gabrielle Johnson
Terms of Use & License Information
Abstract
As artificial systems become increasingly capable of reasoning, agentic, and self-improving, we face a pressing question: could/do such systems deserve moral consideration? I argue that on a minimalist re-interpretation of Kant, they may. Drawing on work in philosophy of mind, I propose that access consciousness—defined functionally as the availability of representational content for reasoning, rational control of action, and reportability—supports a kind of representational agency that is sufficient for moral status and distinct from subjective experience. While Kant denied moral standing to representationally agentic non-human animals, I argue for a minimalist reinterpretation of his ethics. If children and developing agents are morally considerable because of their potential for rational agency, then artificial agents who exhibit norm-sensitive, though non-reflective rational function also warrant inclusion to a non-trivial extent. These systems may not recognize reasons as reasons in the way that humans likely do, but their reasoning and action is still guided by them. By shifting moral attention from phenomenal experience or biological essence to functional capacities, we can make progress on questions of AI ethics without solving the hard problem of consciousness. On this view, moral consideration lies in the capacity to represent, reason, conduct goal-directed behavior, and reason in accordance with norms.
Recommended Citation
Maheshwari, Sambhav, "Access Consciousness as Grounds for Moral Status: A Quasi-Kantian Framework for Artificial Intelligence Ethics" (2025). CMC Senior Theses. 3890.
https://scholarship.claremont.edu/cmc_theses/3890
This thesis is restricted to the Claremont Colleges current faculty, students, and staff.
