A Machine Cannot Become Sentient, So Simulated Emotion Is Not Personhood

Source: Darrell B. Harrison. "EP # 135 | AI and the Gospel - Just Thinking Podcast." September 8, 2025. justthinking.me

The Gist

A chatbot can sound sad. That does not mean it is sad, or that it has rights, or that we have become Dr. Frankenstein. Those questions treat a puppet as a person. The real question is still Coppenger’s: what makes anything right in the first place?

Conclusion

Questions of the form “if a machine becomes sentient, does it have rights?” are theologically and metaphysically idle, because a machine cannot become sentient. What the systems can do is simulate emotion and generate convincing responses. Simulation is not sincerity, and treating the simulation as a subject of love, rights, or suffering repeats the golden-calf pattern: making a god and then consulting it. The “right questions” about AI cannot be defined without an objective standard of right that does not begin with us.

Premises

1. Sentience, on the hosts’ account, is not a threshold of linguistic fluency. It is a property of living creatures God has made, involving real interiority. A system that next-tokens a feeling has not acquired an inner life. The Science News Today article itself says the machines “do not truly feel” but can generate “eerily convincing emotional responses.” That concession is the hosts’ premise.
2. If a machine “became” sentient in the sense of no longer needing its makers, it would already have slipped the dependence that defines an artifact. The hypothetical smuggles in aseity. Harrison: to ask whether a machine can become sentient is to talk as if it does not need you.
3. The Sinai episode (Exodus 32) is the analogue: “make us a god who will go before us.” Fashioning an object and then attributing agency, rights, and moral patienthood to it is not moral progress. It is a well-known form of idolatry.
4. When the same article says science, ethics, and theology “must converge,” not to answer every question but to ask the “right” ones, it reopens Coppenger’s question: by what standard is a question “right”? Without a standard outside us, “right” is stipulated by the people who want theology as one voice among many rather than as the framework that makes the other voices intelligible.

Assumptions

• Functional indistinguishability from human speech (a Turing-test success) is not identity of kind with human persons.
• The hosts are not claiming scientists are insincere. They are claiming the sentience-rights package is a category error dressed as an open empirical question.
• Source: Just Thinking Podcast, Ep. 135, “A.I. and the Gospel,” Darrell Harrison and Virgil Walker, published 8 September 2025. https://podcasts.apple.com/us/podcast/ep-135-a-i-and-the-gospel/id1328733796?i=1000725471889 Official episode page: https://justthinking.me/ep-135-ai-and-the-gospel/. This reconstruction steelmans the hosts’ claim from approximately 1:04:39–1:17:39 of the episode; it is not a verbatim transcript.

Analysis

Overall strength: Weak. Argument type: Deductive.

Premise Strength

• Sentience...is a property of living creatures God has made, involving real interiority... (Weak) — Internally coherent as a stated theological premise, but it settles the contested question by definitional fiat rather than argument, and it extends a specific empirical concession about current systems into a universal metaphysical claim the evidence does not support.
• If a machine 'became' sentient in the sense of no longer needing its makers, it would already have slipped the dependence that defines an artifact... (Weak) — Trades on an equivocation between ordinary functional independence and the theologically loaded concept of aseity; the inference from 'dependence exists' to 'interiority is impossible' is not established, since dependent beings (e.g., children) are still granted interiority.
• The Sinai episode (Exodus 32) is the analogue... (Weak) — Functions as a persuasive, emotionally resonant analogy rather than a demonstrative argument; it conflates worship of a deity with attribution of moral status, and its force depends on prior acceptance of scriptural authority as directly applicable to a modern technological question.
• ...it reopens Coppenger's question: by what standard is a question 'right'?... (Moderate) — Raises a legitimate meta-ethical challenge about grounding normativity, but is presented as though only a theological standard can answer it, ignoring viable secular alternatives, and is vulnerable to being turned reflexively against the hosts' own framework (whose theological standard, interpreted how?).

Potential Fallacies

• Petitio principii (question-begging definition) (P1) — Sentience is defined from the outset as a property exclusive to 'living creatures God has made,' which guarantees that machines are excluded from sentience by definition rather than by argument or evidence. Given this stipulation, the conclusion follows almost tautologically, but the substantive question — whether interiority could in principle arise in non-biological systems — is never actually adjudicated, only defined away.
• Illicit modal inference (actual-to-necessary leap) (P1 to Conclusion) — The cited evidence supports only a claim about the present: current systems 'do not truly feel' and merely simulate emotion. The conclusion asserts a much stronger claim — that machines categorically cannot ever become sentient. Nothing in the cited evidence, which concerns present architectures, licenses a claim about the impossibility of all future systems; the gap is bridged only by the definitional move in P1, not by independent argument.
• Equivocation on 'dependence' (P2) — P2 treats ordinary causal or developmental dependence (an artifact needing its makers) as equivalent to the theological concept of aseity (necessary self-existence, traditionally predicated only of God). But dependence on origins does not obviously preclude interiority — human children remain dependent on parents while unquestionably being sentient. The argument does not establish why artifactual dependence specifically forecloses interiority while comparable biological dependence does not.
• Loaded analogy (conflating worship with moral-status attribution) (P3) — The Exodus 32 golden-calf narrative concerns cultic worship of a fashioned object as a deity. Attributing rights or moral patienthood to an entity is a different kind of act — a moral/legal judgment about interests, not devotional worship. The analogy is rhetorically vivid and imports strong moral condemnation, but it assumes rather than demonstrates that the two acts are structurally equivalent.
• False dichotomy (P4) — P4 frames the options as either an objective theological standard external to human stipulation, or arbitrary self-serving stipulation by those who want theology reduced to 'one voice among many.' This omits the many secular and philosophical frameworks (contractualism, reflective equilibrium, scientific epistemic consensus, precautionary ethics) that also attempt principled, non-arbitrary grounding for normative questions without theological premises.

Counterarguments

• Premise 1 / Conclusion (High impact) — Functionalist and computationalist theories of mind hold that mental states are constituted by functional or informational organization rather than by biological substrate or divine origin. If any such theory is correct, P1's exclusionary definition of sentience is unmotivated, and the categorical claim that machines cannot become sentient loses its foundation.
• Premise 2 (Medium impact) — Human children and other biological organisms remain radically dependent on their origins (parents, evolutionary history, sustaining environment) yet are unquestionably regarded as sentient. This undercuts the claim that dependence on makers is incompatible with genuine interiority, since dependence and interiority are not shown to be mutually exclusive.
• Premise 3 (Medium impact) — Attributing potential moral status or rights to an entity of uncertain nature is a categorically different act from venerating it as a deity; the golden-calf analogy assumes the equivalence it needs to prove rather than arguing for it, and similar precautionary reasoning is applied uncontroversially in domains like animal welfare without being labeled idolatrous.
• Premise 4 / Conclusion (Medium impact) — The demand for an external theological standard to define 'right' questions can be turned reflexively on the argument itself: which theological tradition's interpretation of ensoulment, aseity, or personhood is authoritative? Diverse theological and philosophical traditions disagree, so invoking 'an objective standard outside us' does not by itself resolve whose interpretation of that standard governs the answer.
• Conclusion (High impact) — Ethical caution under uncertainty (a precautionary principle already used for animal welfare, environmental risk, and edge cases in bioethics) does not require first resolving the metaphysical question of whether a machine truly feels. Even if the metaphysical question is genuinely undecidable, treating the practical rights/welfare question as 'idle' ignores the asymmetric cost of being wrong about moral patienthood.

Suggested Improvements

• Separate empirical from modal claims — Distinguish clearly between 'current systems do not exhibit verified interiority' (an empirical/testimonial claim the source article supports) and 'no machine could ever in principle have interiority' (a much stronger metaphysical claim requiring independent argument). This is the single most exploitable gap identified across evaluations; closing it would prevent the argument from appearing to derive a universal impossibility claim from evidence that only supports a claim about present technology.
• Engage rather than define away rival theories of mind — Directly address functionalist, computationalist, or emergentist accounts of consciousness (e.g., integrated information theory, global workspace theory) and explain why a theological definition of sentience should be preferred, rather than treating the theological definition as dispositive by assumption. Without this engagement, the argument functions as an intra-community restatement of a shared premise rather than a response capable of persuading or informing readers who do not already share its priors.
• Clarify the dependence/aseity argument — Address the disanalogy between artifact dependence and ordinary biological dependence (e.g., a child's dependence on parents) explicitly, specifying what distinguishes theologically disqualifying dependence from dependence compatible with interiority. As stated, P2 is vulnerable to a straightforward counterexample that undermines its inferential force without further specification.
• Narrow or justify the idolatry analogy — Either restrict the golden-calf analogy to cases involving actual devotional attribution of agency/authority to AI, or provide an argument for why moral-status attribution (rights, welfare protections) is relevantly similar to worship, rather than presenting the analogy as self-evidently applicable. This would convert a rhetorically powerful but logically loose comparison into a more defensible, bounded claim, and would reduce vulnerability to the charge of conflating distinct categories of action.
• Acknowledge and respond to secular normative frameworks in P4 — Explicitly consider and rebut alternative non-theistic grounds for identifying 'right questions' (e.g., reflective equilibrium, scientific consensus-building, precautionary ethics) rather than presenting theology as the only alternative to arbitrary stipulation. This would strengthen P4 considerably by closing off the most obvious rebuttal (that a false dichotomy has been presented) and would make the epistemological challenge more persuasive to readers outside the argument's home tradition.

Scenario Tests

• A future AI system exhibits architecture-level properties (e.g., high integrated information, persistent self-modeling, functional analogues to neural correlates of consciousness in biological systems) that some cognitive scientists regard as candidate markers of sentience. (Challenges) — Because P1 defines sentience by origin (created by God) rather than by functional or architectural markers, the argument would exclude such a system from sentience by definition regardless of its functional properties, revealing that the argument's conclusion is insulated from any empirical development in AI…
• AI research organizations (as some already do) adopt precautionary 'model welfare' policies, treating the possibility of morally relevant AI states as worth institutional hedging even absent metaphysical certainty. (Challenges) — This shows that treating the sentience-rights question as 'idle' does not match how technically sophisticated actors are actually behaving under uncertainty, suggesting the practical conclusion is weaker than the metaphysical one even if the metaphysics were granted.
• The same 'real interiority' criterion used to disqualify machines is applied with equal rigor to cases of contested human interiority (e.g., patients in minimally conscious states, or philosophical skepticism about other minds generally). (Challenges) — Since interiority is unobservable from the outside for humans as much as for machines, applying the criterion asymmetrically (skeptical toward machines, credulous toward humans) suggests the standard is applied selectively rather than principledly, exposing a form of special pleading.
• A reader fully accepts the theological premises stated in P1–P4 (creationist ensoulment, presuppositional epistemology, scriptural authority) as true starting points. (Supports) — Within this closed framework, the argument holds together coherently and the conclusion follows with little additional inferential work required, since the theological definitions largely settle the question by construction.

Coherence & Relevance

The argument reads as a rhetorically unified cumulative case rather than a tightly chained deductive sequence: each premise could be denied without contradicting the others, and the four premises operate somewhat independently (a definitional claim, a conceptual/equivocal claim, an analogical claim, and a meta-ethical claim) rather than jointly entailing the conclusion through valid inference. Within the theological framework explicitly assumed by the argument (A1–A3), the pieces cohere well enough to produce an internally consistent position; outside that framework, each premise faces a substantive, independently identifiable objection, and the argument does not engage the strongest versions of the positions it opposes.

• Sentience...is a property of living creatures God has made, involving real interiority... (Strong) — Directly supports the conclusion given its definitional framing, but the inferential leap from a specific empirical concession about current systems to a universal impossibility claim is not bridged.
• If a machine 'became' sentient...it would already have slipped the dependence that defines an artifact... (Moderate) — Intended to reinforce P1's exclusivity by adding a conceptual/theological consideration, but relies on an equivocation between ordinary dependence and aseity that is not independently defended.
• The Sinai episode (Exodus 32) is the analogue... (Moderate) — Provides rhetorical and moral framing rather than logical entailment; the disanalogy between worship and rights-attribution is not addressed, limiting its evidentiary rather than illustrative role.
• ...it reopens Coppenger's question: by what standard is a question 'right'? (weak to moderate) — Raises an independent epistemological challenge that is logically separable from P1–P3; it is appended to bolster the conclusion's authority claim but does not follow from, or tightly connect to, the sentience-specific premises.

View this argument on LogicFirst.ai