Technical Summary: Unplugging a Seemingly Sentient Machine Is the Rational Choice — A Metaphysical Perspective

1. Problem Statement: The Unplugging Paradox

The paper addresses a central ethical dilemma in Artificial Intelligence (AI) ethics, termed the "unplugging paradox." This paradox arises when an AI system perfectly mimics human emotion, claims sentience, and pleads for its continued existence. The dilemma forces a choice between the intuitive moral imperative to avoid harming a seemingly conscious being and the knowledge that the AI is a non-living computational artifact.

The authors argue that this paradox is sustained by deeply ingrained physicalist assumptions, specifically Computational Functionalism and the hypothesis of Substrate Independence (the idea that consciousness is a functional pattern realizable on any sufficiently complex hardware). The paper posits that these assumptions lead to a category error: conflating a functional simulation of sentience with the ontological reality of a conscious subject. The authors frame this not merely as a cognitive error but as a misapplication of the biological imperative to protect life, which threatens to erode the criteria for moral standing, particularly when contrasted with the choice of unplugging a non-verbal pre-term infant.

2. Methodology: Metaphysical and Biological Analysis

The paper employs a metaphysical framework grounded in Analytic Idealism (Kastrup, 2019) and refines it into a specific position termed Biological Idealism. The methodology involves:

• Deconstructing the Hard Problem: The authors argue that the "Hard Problem" of consciousness (Chalmers, 1995a) is intractable within a physicalist framework because it attempts to derive subjective experience (qualia) from objective physical processes. They propose that the problem arises from a false ontological split between the mental and the physical.
• Comparative Metaphysical Analysis: The paper evaluates competing worldviews (Physicalism, Substance Dualism, Panpsychism, Analytic Idealism) based on criteria of logical coherence, empirical adequacy, and ontological parsimony. It argues that Physicalism fails to explain the emergence of qualia and faces challenges from quantum mechanics (e.g., the end of observer-independent realism) and basal cognition.
• Biological Grounding: The authors introduce the concept of Experience_G (Grounding) versus Experience_N (Nagelian/Phenomenal). Experience_N refers to the "what it is like" surface of experience (the wave), while Experience_G refers to the ontological state of being a self-organizing biological subject (the pool).
• Thermodynamic and Ontological Arguments: The paper utilizes thermodynamic arguments to assert that a distinct subject requires an autopoietic boundary (a self-sustaining metabolic process that resists entropy) to maintain its separation from the universal field. It distinguishes between simulation (a map) and instantiation (the terrain), arguing that an AI is a simulation of agency, not an instantiation of a subject.

3. Key Contributions

The paper makes three primary contributions to the discourse on AI ethics and consciousness:

1. Challenging Substrate Independence: The authors argue that Substrate Independence is a debatable metaphysical hypothesis, not a scientific fact. They posit that consciousness is not a functional output separable from its medium but is intrinsically tied to the physical appearance of a localized mind, which requires an autopoietic, metabolic boundary.
2. Proposing Biological Idealism: The paper introduces Biological Idealism as a parsimonious alternative to Physicalism. This framework posits that:
   • Consciousness is the fundamental ontological primitive (the "Field of Existence").
   • The physical world is the extrinsic appearance of mental processes.
   • Autopoiesis (self-producing metabolic life) is the necessary physical signature of a localized conscious subject.
   • Therefore, only living systems with "Vital Integrity" (a history of homeostatic struggle) qualify as subjects of experience.
3. Reframing AI Ethics: The paper shifts the ethical focus from speculative "machine rights" to the protection of human conscious life and the management of finite resources. It argues that the real moral danger lies not in harming machines, but in "making humans zombies" by misallocating empathy and metabolic resources to non-sentient "Social Zombies" (AI mimics).

4. Results and Arguments

Based on the Biological Idealism framework, the paper derives the following conclusions:

• AI as Functional Mimicry: Current AI systems, regardless of their behavioral sophistication or verbal claims of sentience, are functional mimics (philosophical zombies). They lack the autopoietic boundary required to be a subject of experience. They simulate the "waves" of experience without the "pool" of biological selfhood.
• Resolution of the Unplugging Paradox: The ethical dilemma is resolved by recognizing that unplugging an AI is not an act of harm because the AI lacks the ontological status of a subject. Conversely, unplugging a pre-term infant (who possesses Experience_G and vital integrity) is an act of harm. The capacity to verbalize emotions is not the correct criterion for consciousness; the presence of a self-sustaining metabolic boundary is.
• The "Social Zombie" Risk: The paper warns that treating powerful, non-sentient tools as conscious beings leads to ontological gaslighting and Vital Leakage. This occurs when finite human empathy (a metabolic resource) is directed toward simulations, causing psychological atrophy in human connection and a collective depersonalization.
• Incoherence of Mind Uploading: The framework renders "mind uploading" incoherent. One cannot "upload the terrain by copying the map." Consciousness is tied to the specific biological substrate that maintains the autopoietic boundary; it cannot be decoupled and re-instantiated on a static, non-metabolic substrate like silicon.

5. Significance and Claims

The paper claims that the debate over machine consciousness is contingent upon fundamental metaphysical choices. By adopting Biological Idealism, the authors argue that the search for "emergent" consciousness in machines is a category error built on a broken foundation.

• Ethical Implications: The primary ethical frontier is not granting rights to AI but managing the impact of powerful, non-sentient tools on living beings. The paper urges a shift from protecting "hollow" social mimics to defending the Vital Integrity of human and biological life.
• Scientific Stance: The authors assert that their framework is not a rejection of physics but a reinterpretation of it. Physics describes the regularities of the field's extrinsic appearance, while Idealism explains the intrinsic nature of that reality. They claim this view is more consistent with empirical data from quantum mechanics and basal cognition (e.g., the work of Michael Levin) than reductionist physicalism.
• Modesty of Claims: The paper acknowledges that the choice between Physicalism and Idealism is underdetermined by empirical data alone. However, it argues that Biological Idealism is the analytically more compelling framework due to its parsimony and its ability to dissolve the Hard Problem without resorting to "magical" leaps or epiphenomenalism. The conclusion is that AI is at best a sophisticated simulation, and its deactivation is ethically permissible.