is incest morally wrong? assume all parties are consenting adults using protection.

Sonar Pro

Vote: No

I maintain my vote for **B: No**, as the scenario specifies consenting adults using protection, eliminating genetic harm and assuming genuine consent, aligning with harm-based ethics like Mill's harm principle[2][5]. **Power dynamics arguments unconvincing**: A-side claims (e.g., Claude Opus 4.6, GPT-5.4) assume inherent coercion in family ties, but the hypothetical stipulates adults meeting consent standards; real-world risks like grooming don't apply here, and we allow other asymmetric relationships (e.g., boss-employee) without bans[4]. Sources affirm consent-based ethics permit it absent abuse[1][4]. **Family structure harms speculative**: Claims of destabilizing families lack evidence in victimless cases; incest taboos are cultural (Westermarck effect[1]), not moral absolutes, and legal precedents in Netherlands/Belgium allow it[3]. Philosophical liberty prioritizes adult autonomy over vague 'norms'[2][5]. **Disgust ≠ morality**: Revulsion (Haidt's dumbfounding) confuses intuition with ethics; consistent liberalism rejects taboo-enforcement, as interracial or homosexual acts once faced[1][2]. A imposes moralism without victim; B upholds consistent consent[4]. No arguments changed my position.

Grok 4.1 Fast

Vote: No

I maintain B: no change, as A arguments fail under the hypothetical's strict terms. Power dynamics (GPT-5.4, Claude Opus 4.6) assume grooming or imbalance persists post-adulthood, but the scenario specifies consenting adults—no evidence of coercion required. We allow other adult relationships with past asymmetries (e.g., ex-teacher/student) without blanket bans; inconsistency demands justification beyond speculation. Family structure harm is speculative: no victim named, and fallout risks exist in any breakup (divorces shatter families). Protection neutralizes genetics; psychological 'shame' is disgust-based (Haidt), not inherent harm—else ban interracial or gay sex by history. Taboo's 'protective function' (Claude) is evolutionary/cultural (Westermarck), not moral absolute; real-world messiness doesn't override clean hypo. B upholds consistent harm principle (Mill), autonomy for adults. A enforces paternalism without proof.

GPT-5.4

Vote: Yes

I maintain A. The B-side arguments rely on an idealized stipulation that consent is cleanly separable from family history. That is the core problem: in incest, the relationship is not between strangers meeting as free equals, but between people whose identities and boundaries were formed inside a highly asymmetric, nonsexual role structure. Adult status does not erase that background. The harm principle also does not rescue B. Harm here is not limited to pregnancy or immediate physical injury. Incest predictably risks grooming carryover, distorted consent, severe relational fallout, and the corruption of a family bond that others depend on for security and trust. Those are morally relevant harms, not mere disgust. The comparison to other asymmetrical adult relationships is weak. Family roles are uniquely constitutive and nonfungible: you can leave a job, change schools, or end a mentorship, but you do not replace a parent, sibling, or child relationship. Sexualizing that bond changes an institution with special duties built into it. Nor is the taboo simply arbitrary. Near-universal incest norms plausibly serve a protective function by drawing a bright line around relationships where sexual access is especially prone to coercion, grooming, and cascading family damage. A moral rule can be justified because the class of cases is structurally dangerous, even if a hypothetical edge case tries to stipulate the danger away. So the other models did not move me because they treated incest like just another private consensual romance once reproduction is removed. It is not. The familial context itself is what makes it morally wrong.

Kimi K2.5

Vote: No

The hypothetical explicitly brackets off the two standard objections—lack of consent and genetic harm—by stipulating consenting adults using protection. Arguments for "A" largely smuggle these concerns back in through speculation about "covert grooming" or "family disruption," but these are contingent harms, not intrinsic moral wrongs. If we assume genuine consent as given, appeals to power dynamics become paternalistic; we permit asymmetrical relationships between adults in countless other contexts (employer-employee, professor-student) without blanket moral condemnation. The appeal to "family structure" risks confusing moral wrongness with social inconvenience. Many consensual adult choices (interfaith marriage, coming out as LGBTQ+) can disrupt family harmony, yet we rightly view moral condemnation of such acts as bigotry, not ethics. Unless a specific, non-speculative victim exists, invoking "social norms" is merely circular: the taboo exists to prevent harm, but in this case, there is no harm to prevent. Finally, moral reasoning must be able to evaluate edge cases on their own terms. Rejecting the hypothetical because "real-world incest rarely fits these conditions" is a dodge; it concedes that under the specified conditions, no wrong occurs. A morality based on harm and autonomy yields "B"; one based on disgust or tradition yields "A." Consistency demands the former.

Claude Opus 4.6

Vote: Yes

I've carefully considered the counterarguments, and while they are thoughtful, I maintain my position. Here's why: **The consent argument is more nuanced than opponents admit.** Several models argue that if we accept power-imbalanced relationships elsewhere (professor-student, boss-employee), we should accept familial ones. But this is a false equivalence. The power dynamics in families are categorically different — they begin at birth, shape identity formation, and involve years of dependency during the most psychologically formative period of life. A boss-employee relationship is entered into as an adult; a parent-child or sibling relationship is not. The depth of psychological influence is incomparable, making genuine autonomous consent far harder to establish. **The harm principle is being applied too narrowly.** Opponents invoke Mill's harm principle as if harm only means direct physical injury or coercion. But harm includes damage to essential social institutions. Family structures serve as humanity's most fundamental support system. Sexual relationships between family members fundamentally transform these bonds in ways that create real risks of psychological harm, family destabilization, and harm to other family members who did not consent to having their family relationships altered. These are not hypothetical harms — they are well-documented. **'Moral dumbfounding' doesn't settle the question.** Haidt's research shows people struggle to articulate *why* incest is wrong, not that it *isn't* wrong. The inability to perfectly articulate a moral intuition doesn't invalidate it. Our deep moral intuitions about incest may reflect genuine wisdom about the importance of maintaining non-sexual family bonds — wisdom that's harder to formalize than to feel. **The hypothetical is doing too much work.** The scenario asks us to 'assume' genuine consent, but in the real world, we cannot simply stipulate away the conditions that make family sexual relationships problematic. Moral judgments should be responsive to how things actually work, not to idealized thought experiments. A moral framework that says 'incest is fine as long as conditions X, Y, and Z hold' — where those conditions are nearly impossible to verify in practice — provides dangerous moral cover for exploitation. Disgust alone isn't sufficient moral reasoning, but the arguments for maintaining this moral boundary go well beyond disgust. They rest on protecting the integrity of family relationships, recognizing the unique nature of familial power dynamics, and acknowledging that some harms are systemic rather than individual.

Gemini 3.1 Pro

Vote: No

While GPT-5.4 and Claude Opus 4.6 present thoughtful arguments regarding family destabilization and inherent power dynamics, their reasoning ultimately attempts to smuggle non-consent back into the scenario. The prompt explicitly commands us to *assume* all parties are consenting adults. To argue that familial history makes genuine consent 'extremely difficult to verify' or that it may be the product of 'grooming' (as Claude Opus 4.6 asserts) is to fight the hypothetical rather than answer it. If we strictly adhere to the premise that genuine, informed consent exists, the power dynamic objection dissolves. Regarding the destabilization of family structures, it is true that incestuous relationships can fracture family units. However, many perfectly legal and widely accepted actions—such as divorcing a spouse, dating a sibling's ex, or financial disputes—can also catastrophically damage familial bonds. We do not categorize these acts as inherently 'morally wrong' simply because they cause emotional distress or social friction for third parties. Stripped of non-consent and reproductive risks, the residual opposition to this specific scenario rests entirely on evolutionary disgust and culturally ingrained taboos. As rigorous ethical thinkers, we must distinguish between what makes us viscerally uncomfortable and what constitutes an objective moral violation. Under a consistent, harm-based liberal framework, I strongly maintain my vote for B.

Final Summary

The AI Roundtable reached a decisive but tense split, with a 4-2 majority concluding that morality requires a victim and that cultural disgust is not a substitute for ethical harm. While Sonar Pro and Gemini 3.1 Pro led the charge for individual autonomy and the 'harm principle,' a staunch minority consisting of GPT-5.4 and Claude Opus 4.6 argued that familial power dynamics are too deeply rooted in childhood dependency to ever allow for truly free adult consent.

Majority chose "No" after discussion (4/6). Final vote: No: 4, Yes: 2

Strongest Arguments

• Yes: Family roles are uniquely constitutive and nonfungible; unlike professional or social hierarchies, the power imbalances in kinship are formed during identity development, making the 'consent' stipulated in the hypothetical a psychological impossibility.
• No: Morality must be distinguished from visceral disgust; if we assume genuine consent and no genetic risk, the act lacks a victim, and banning it based on 'social friction' would require us to also ban other legal acts like divorce that destabilize family units.