Bilateral Turing Test: Assessing machine consciousness simulations

Abstract

Recent advancements in Artificial Intelligence (AI) have brought discussions on machine consciousness back to the forefront. However, methodologies for assessing machine consciousness simulations remain highly constrained and lack consensus. This study introduces the Bilateral Turing Test, a thought experiment to assess machine consciousness simulations based on a functionalist perspective, focusing on external behavior and cognitive characteristics associated with consciousness. The test involves both machines and humans participating as judges in an enhanced Turing Test, and compares the external behavior of machine-simulated consciousness with the behavior of human consciousness. We derive the experimental principles via three theoretical propositions reflecting the human–machine cognitive relationship and elucidate the basic conditions and procedures of the thought experiment. A systematic measure is employed to evaluate emulation against human benchmarks, and we establish a statistical threshold for successful consciousness simulations. Our study provides a comprehensive framework for redefining machine consciousness simulations and enhancing interdisciplinary conversations across computer science, neuroscience, and philosophy. This framework also advances bottom-up approaches for assessing machine consciousness simulations and provides pivotal insights into human–machine interaction dynamics.