On Possible Advantages of Conscious Systems and a Turing Test for Consciousness
Advances in Artificial Intelligence and Machine Learning 2024 DOI: 10.54364/aaiml.2024.44183 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper Peer reviewed |
|---|---|
| Key points | Argues that systems with inherent quantum processes could form "hypercomplex system states" that store information in unmeasurable energy states and could aggregate into consciousness, and proposes that technical systems able to use such states would differ in performance from those that cannot, offering a "Turing test for consciousness" and a design for initial experiments to evaluate these mathematically derived predictions. |
Abstract
Earlier articles have suggested that physical systems with inherent quantum processes could be capable of producing hypercomplex system states. These system states could have special energy states, which are referred to as hypercomplex energy states because they cannot be measured, although they physically exist and could even store information. Hypercomplex system states could be aggregated into consciousness in certain systems, such as the human brain. But technical systems could also potentially generate and use hypercomplex system states, which would create a new level of AI systems. Systems capable of consciousness must and will differ significantly in their performance from systems that cannot utilize such states. This performance difference could be used to distinguish systems with technical consciousness (so-called ”machine consciousness”) from systems without ”consciousness”. Based on this, the article presents a so-called Turing test for consciousness for technical systems, in particular AI systems. The hypotheses in this and previous articles are predictions that arise from mathematical considerations, but they still have to be subjected to experimental evaluation. This article presents the design for initial experiments.