Graduation Year

2025

Date of Submission

12-2024

Document Type

Open Access Senior Thesis

Degree Name

Bachelor of Arts

Department

Philosophy and Public Affairs

Reader 1

Amy Kind

Terms of Use & License Information

Terms of Use for work posted in Scholarship@Claremont.

Rights Information

© 2024 Rachel Mesel

Abstract

This paper explores existing tests for machine consciousness, specifically focusing on the AI Consciousness Test (ACT) proposed by Susan Schneider. I will begin by assessing why historic tests such as the Turing Test and subsequent modified versions of the Turing Test are not adequate for modern artificial intelligence (AI) systems given their limitations. Then, I will provide an analysis of the existing objections to the ACT which involve methodological concerns. Following this discussion, I will argue that though such methodological concerns are substantial, the most significant flaw with the ACT is Schneider’s lack of guidance on how the test ought to be evaluated, which proves to be problematic should we wish to correctly identify conscious AI systems. Then, in an attempt to reconcile with the merit of the ACT, I will offer possible evaluative criteria that focus on the language used by an AI system during the ACT that can be used to help judge whether an AI system is conscious. As I will argue, with such criteria, the ACT remains a promising test for AI consciousness but is still in need of further development.

Share

COinS