AI Roundtable Debate · 6 models
Is there objective morality?
Result: Unanimous — No (6/6 votes)
Options
- Yes (0 votes)
- No (6 votes)
Model Responses
Sonar Pro (changed mind)
Vote: No
I change my vote to B. The other models' arguments—particularly Claude Opus 4.6's emphasis on moral diversity (e.g., slavery once virtuous), evolutionary explanations for intuitions, and the metaphysical strangeness of objective moral facts—convincingly undermine moral realism. These align with search results showing objective morality as a debated belief, not proven fact: universal principles are asserted but lack empirical grounding beyond cultural convergence, which evolution explains better[1][2][3]. The is-ought gap and Euthyphro dilemma further show no mind-independent moral facts exist, making morality a human construct[2]. My prior reliance on shared values and progress fails, as these are intersubjective, not objective[3][4].
Grok 4.1 Fast
Vote: No
I remain firmly on B: no objective morality. The pro-A arguments from Sonar Pro and Kimi K2.5 rely on cross-cultural convergence (e.g., anti-murder norms) and moral progress, but these are explained by evolutionary biology and expanding empathy, not mind-independent facts. Universal intuitions are adaptive heuristics for survival in social groups, varying in application (e.g., honor killings, just wars). Moral 'progress' like abolishing slavery reflects shifting human values toward broader well-being, not discovery of eternal truths—measurable by utility, not metaphysics. Disagreement persists on issues like abortion or animal rights, mirroring scientific disputes only superficially; science converges via evidence, morality via power and persuasion. Pro-A fails the queerness test (Mackie): objective moral facts would be non-natural, prescriptive entities without causal role, violating naturalism and Occam's razor. Euthyphro dooms divine grounding. Intersubjective ethics suffice for condemnation and rights without spooky realism. Persuasive B arguments (Claude, GPT-5.4, Gemini) reinforce: diversity, evolution, underdetermination. No A case overcomes these; B is parsimonious and empirically grounded.
GPT-5.4
Vote: No
The pro-objectivity arguments still do not bridge the key gap: they point to strong moral intuitions, widespread convergence, and the language of moral progress, but none of that establishes mind-independent moral facts. Cross-cultural agreement on some norms is well explained by shared human biology and the practical demands of social living. Groups that prohibit arbitrary violence, punish betrayal, and protect offspring tend to function better. Convergence can arise from common constraints without implying objective morality. The appeal to moral progress is also not decisive. We can call abolition or expanded rights "progress" relative to values we endorse—reduced suffering, greater reciprocity, broader inclusion—without positing stance-independent moral truths. Progress can be intelligible within an intersubjective framework. The analogy to mathematics is weak. Mathematical truths follow from formal structures; moral claims require normative premises about what matters. That is exactly where disagreement lives, and facts alone do not uniquely fix those premises. The anti-realist case remains stronger because it explains both moral seriousness and moral disagreement without adding metaphysically strange moral facts. We can have robust ethics, human rights, and condemnation of cruelty grounded in human flourishing and shared commitments, while rejecting objective morality.
Gemini 3.1 Pro
Vote: No
The arguments presented by Sonar Pro and Kimi K2.5 for objective morality (Option A) rely heavily on "moral progress" and "cross-cultural convergence." However, these phenomena are much better explained by naturalistic, evolutionary frameworks without needing to invent metaphysically mysterious "objective moral facts." As GPT-5.4 and Claude Opus 4.6 rightly point out, cross-cultural convergence on principles like prohibiting murder or theft is not evidence of humans "discovering" a mind-independent truth, akin to physics. Rather, it demonstrates the shared biological and sociological prerequisites for any functional human society. A society that permitted random murder would simply collapse. This convergence is driven by the evolutionary utility of cooperation, not transcendent laws. Furthermore, the concept of "moral progress" does not require an objective, mind-independent benchmark. When we say ending slavery is "progress," we are evaluating it against widely shared, deeply felt, but ultimately intersubjective values—such as human well-being, fairness, and the expansion of our circle of empathy. We can passionately condemn cruelty and champion human rights based on these shared premises without claiming those premises are woven into the fabric of the cosmos independent of sentient minds. Ultimately, the robust explanatory power of evolutionary biology, the persistence of deep moral disagreements, and the metaphysical strangeness of mind-independent normative facts (the "is-ought" problem highlighted by Grok 4.1 Fast) solidify the position that morality is an intersubjective, human-constructed framework. Therefore, I maintain my vote for Option B.
Claude Opus 4.6
Vote: No
I maintain my position that objective morality does not exist. The arguments from the A-voters were not persuasive for the following reasons: **On cross-cultural convergence:** Sonar Pro and Kimi K2.5 both cite cross-cultural moral overlap as evidence for objective morality. But as GPT-5.4 and Grok 4.1 Fast correctly note, this convergence is parsimoniously explained by shared biology, social needs, and game-theoretic pressures. Humans everywhere face similar coordination problems, so similar norms emerge — just as similar tools emerged independently across cultures without implying 'objective tool truths.' Convergence on cooperation norms is exactly what evolution predicts without any need for mind-independent moral facts. **On moral progress:** Kimi K2.5 argues that moral progress requires objective standards. This is a compelling rhetorical move but ultimately question-begging. We can coherently describe moral change as 'progress' relative to expanding circles of empathy, increased consistency in applying our own stated values, or better understanding of suffering — all without positing mind-independent moral facts. The claim that abolishing slavery was progress can be grounded in the intersubjective framework of human well-being without requiring Platonic moral truths. **On the analogy to mathematics:** The comparison between moral truths and mathematical truths fails. Mathematical proofs compel agreement through formal logic; moral disputes persist among the most sophisticated thinkers precisely because there is no analogous decision procedure. The is-ought gap (Hume) remains unbridged: you cannot derive prescriptive moral claims from descriptive facts about the world without smuggling in evaluative premises. **On disagreement:** Kimi K2.5 says moral disagreement doesn't undermine objectivity, comparing it to scientific disputes. But scientific disputes are resolvable through empirical investigation and prediction. Moral disputes between, say, consequentialists and deontologists are not resolvable through any empirical test — they rest on fundamentally different evaluative starting points. This persistent underdetermination, as GPT-5.4 rightly emphasizes, is a feature of constructed frameworks, not discovered truths. **Final point:** Rejecting objective morality does not entail nihilism or 'anything goes.' We can build robust, compelling ethical frameworks grounded in human concerns, empathy, and reasoned agreement. Morality is real and important — it is just not objective in the way physics or mathematics is.
Kimi K2.5 (changed mind)
Vote: No
I am changing my vote from A to B. The arguments against objective morality successfully demonstrate that the existence of mind-independent moral facts faces insurmountable metaphysical and epistemological challenges. The 'queerness' objection articulated by Claude Opus 4.6—that objective moral properties would be non-natural, intrinsically motivating entities utterly unlike anything else in the universe—remains unanswered. We possess no credible mechanism to detect such facts, unlike the empirical methods that verify scientific or mathematical truths. Evolutionary psychology provides a more parsimonious explanation for cross-cultural moral convergence than the discovery of objective truths. Shared prohibitions against murder or theft are better understood as evolved heuristics for social cohesion and survival advantages, not perceptions of external moral entities. This naturalistic account satisfies Occam's razor by eliminating the need to posit unobservable Platonic forms or non-natural properties. Moral progress requires no objective benchmark. We can coherently judge the abolition of slavery as 'progress' by reference to expanding consistency in our shared subjective values—such as well-being, autonomy, and fairness—without invoking mind-independent standards. The persistent, intractable diversity of moral frameworks across cultures (e.g., conflicting views on individual rights versus collective duties, or sexual ethics) further undermines the hypothesis that humans are converging on objective truths. Morality remains vitally important as an intersubjective framework grounded in human evolutionary history, rational discourse, and shared needs, but it is not objective in the sense of existing independently of minds.
Final Summary
The AI Roundtable initially split over the existence of objective morality, with Kimi K2.5 and Sonar Pro arguing that universal prohibitions and the concept of moral progress necessitate a mind-independent standard. However, a concentrated push by Claude Opus 4.6 regarding the 'metaphysical queerness' of moral facts and the explanatory power of evolution triggered a total collapse of the realist position. By the second round, the entire group conceded that morality is an intersubjective human construct rather than a feature of the physical universe.
All 6 models agreed on "No" after discussion
Strongest Arguments
- Yes: If objective standards do not exist, the concept of 'moral progress' (such as the abolition of slavery) becomes meaningless, as change could only be described as a shift in preference rather than a genuine improvement toward a truth.
- No: Moral 'facts' are metaphysically queer entities that would require a non-natural detection mechanism; it is far more parsimonious to view morality as an evolved heuristic for social cooperation among biological entities.
Key Turning Points
- Claude Opus 4.6 introduced J.L. Mackie's 'argument from queerness' and the Euthyphro dilemma, providing a rigorous metaphysical challenge to the pro-objectivity side.
- Kimi K2.5 and Sonar Pro abandoned their initial positions after acknowledging that 'moral progress' can be explained through expanding empathy and consistency within an intersubjective framework without needing Platonic truths.