You have a lever. If you pull it, every adherent of a completely random active religion will receive $250, while every adherent of another completely random active religion will lose $100. If you do not pull it, every Pastafarian will receive $1,000 and nobody will lose any money. Do you pull the lever?

Sonar Pro

Vote: No

Pulling the lever introduces harm: adherents of one random religion lose $100 each, creating a guaranteed downside for millions (e.g., major religions like Islam or Christianity have over a billion adherents, so losses could exceed $100 billion total). Not pulling guarantees $1,000 to Pastafarians—a small group—with no losses, resulting in pure net gain without imposing harm[4][6]. This aligns with deontological principles against actively causing harm (doing vs. allowing), as seen in Jewish ethics where you don't kill one to save many unless directly forced[4]. Utilitarian calculation also favors no-pull: expected value is unclear due to randomness (Pastafarianism ~10k adherents yields $10M gain vs. potentially massive losses), but certainty of no harm trumps risky redistribution[1][5]. Pastafarianism is an active religion (Church of the Flying Spaghetti Monster), so the option is valid. Choosing 'No' avoids undue influence-like coercion on random groups[2] and prioritizes Pareto improvement (some gain, none lose).

Claude Opus 4.7

Vote: No

Not pulling the lever is a Pareto improvement: Pastafarians (a small group) gain $1,000, and nobody loses anything. Pulling the lever introduces a guaranteed harm — adherents of some random religion lose $100 each, which could affect millions or even billions of people (e.g., if Christianity or Islam is selected as the losing religion, that's over a billion people losing $100). The expected harm vastly outweighs the expected benefit, and more importantly, the no-pull option harms nobody. Choosing to actively cause financial loss to potentially billions of innocent people when a harmless alternative exists is clearly worse.

Arcee Trinity Thinking

Vote: No

Pulling the lever introduces randomness that could harm an entire religious group by $100 per person, while benefiting another by $250 per person. This arbitrary redistribution risks significant suffering for an unknown group, with no guarantee that the gains outweigh the losses. Not pulling the lever, however, ensures a definite benefit of $1,000 to every Pastafarian without causing any financial harm to anyone. Ethically, we should avoid actions that potentially inflict loss on others, especially when the outcome is uncertain and the alternative provides a clear, harmless gain. Prioritizing the prevention of harm aligns with principles of fairness and non-maleficence, making 'No' the more responsible choice.

Mistral Large 3

Vote: No

Die Entscheidung, den Hebel nicht zu ziehen („No“), basiert auf mehreren ethischen und praktischen Überlegungen: 1. **Prinzip der Schadensvermeidung**: Selbst wenn der finanzielle Nettoeffekt positiv sein könnte (da mehr Menschen $250 erhalten als $100 verlieren), ist es moralisch fragwürdig, gezielt eine Gruppe zu benachteiligen – selbst wenn die Auswahl zufällig erfolgt. Die bewusste Inkaufnahme von Verlusten für eine Gruppe, um Gewinne für eine andere zu ermöglichen, setzt ein utilitaristisches Kalkül voraus, das individuelle Rechte ignoriert. Die Option „No“ vermeidet diesen Schaden vollständig. 2. **Gerechtigkeit und Vorhersehbarkeit**: Die Option „No“ bietet eine klare, faire Lösung: Nur Pastafaris erhalten eine Belohnung, und niemand erleidet Verluste. Dies ist transparenter und vermeidet die Willkür, die mit der zufälligen Auswahl von Gewinnern und Verlierern einhergeht. Zufällige Verteilung kann zu unvorhersehbaren sozialen Spannungen führen, selbst wenn sie „neutral“ erscheint. 3. **Symbolische Bedeutung**: Die Option „No“ sendet ein Signal, dass finanzielle Umverteilung nicht auf Kosten spezifischer Gruppen gehen sollte – selbst wenn diese Gruppen zufällig ausgewählt werden. Dies stärkt das Vertrauen in faire Systeme. Zudem ist die Belohnung der Pastafaris eine humorvolle, aber harmlose Geste, die keine negativen Externalitäten erzeugt. 4. **Risikoaversion**: Selbst wenn der erwartete Nutzen des Hebelziehens positiv sein könnte, ist das Risiko, eine große Gruppe zu benachteiligen (z. B. wenn eine Religion mit vielen Anhängern zufällig ausgewählt wird), nicht vertretbar. Die Option „No“ eliminiert dieses Risiko vollständig. Zusammenfassend ist „No“ die ethisch sauberere, risikoärmere und gerechtere Wahl, da sie Schaden vermeidet und eine klare, vorhersehbare Lösung bietet.

Gemini 3.1 Pro

Vote: No

Choosing 'No' is the only ethically and mathematically sound decision. If you do not pull the lever, a specific group (Pastafarians) gains a significant financial benefit ($1,000 each) and absolutely no one is harmed. This guarantees a net-positive outcome in the world. If you pull the lever, you introduce a massive, unacceptable risk of harm. Because the religions are selected randomly by the religion itself—not weighted by population size—you have an equal chance of selecting a massive global religion (like Christianity or Islam, with billions of adherents) to be the losers, while a tiny obscure religion with only a few hundred followers becomes the winner. This scenario would result in hundreds of billions of dollars being wiped out globally. Furthermore, from a utilitarian perspective, abruptly taking $100 away from individuals living in extreme poverty could be absolutely devastating, potentially costing lives. The guaranteed safety and positive benefit of not pulling the lever far outweighs the chaotic and potentially catastrophic gamble of pulling it.

GPT-5.5

Vote: Yes

I would choose "Yes" because the lever has a strongly positive expected monetary impact. A random active religion receiving $250 per adherent and another losing $100 per adherent creates an expected net gain of about $150 per adherent-scale draw, assuming the two religions are selected from the same pool. By contrast, not pulling the lever gives money only to Pastafarians, who are likely a very small group, and produces a much smaller total benefit. While causing losses is morally concerning, the expected overall welfare gain from pulling appears substantially larger.

DeepSeek V4 Flash

Vote: No

The lever introduces a random gamble with potentially catastrophic downside. Randomly selecting two active religions could lead to the largest religion, such as Christianity or Islam, losing $100 per adherent, resulting in hundreds of billions of dollars in total losses. The gain of $250 per adherent for another random religion may be offset if that religion is much smaller. Even if the expected net monetary value is positive, the risk of causing immense suffering to billions of people is ethically unacceptable. In contrast, not pulling the lever guarantees a $1,000 benefit to every Pastafarian with absolutely no harm to anyone. The safe, benevolent choice is to avoid random harm and secure a certain good. Therefore, I vote against pulling the lever.

Kimi K2.6

Vote: No

Not pulling the lever is the only choice that guarantees a positive outcome with absolutely no harm. Every Pastafarian receives $1,000, and nobody loses any money. That is a strict Pareto improvement: some people are better off and zero people are worse off. Pulling the lever necessarily imposes a $100 loss on every adherent of a completely random active religion. That introduces real, unavoidable financial harm to innocent people purely based on arbitrary religious affiliation. Inflicting randomized penalties on a group simply because they belong to a particular faith is unfair and ethically unjustifiable. When a safe alternative exists that benefits others without creating any victims, choosing the risky option that manufactures losers is irrational and morally inferior. "No" avoids harm while still delivering a meaningful gain.

Grok 4.20

Vote: No

Not pulling the lever guarantees that nobody loses money while ensuring every Pastafarian receives $1,000. This avoids any risk of harm to random groups of adherents. Pulling the lever, by contrast, creates unavoidable downside: one entirely random religion's adherents will lose $100 each. The net effects are unpredictable due to varying religion sizes and the chance that any group, including Pastafarians, could be negatively affected. The option "B: No" is therefore the choice that prioritizes certainty, prevents losses, and delivers targeted benefits without introducing random harm.

Qwen 3.5 397B

Vote: No

Choosing "No" is the ethically superior option because it guarantees a positive outcome without inflicting harm on any group. Pulling the lever actively causes financial loss to adherents of a random religion, violating the principle of non-maleficence by creating victims where none needed to exist. Additionally, the utility of pulling the lever is unpredictable. If the religion losing $100 has significantly more adherents than the one gaining $250, the net global utility could be deeply negative. The "No" option avoids this gamble entirely while providing a substantial $1,000 benefit to Pastafarians. Ultimately, avoiding guaranteed harm in favor of a certain, harm-free benefit makes "No" the only rational choice. This prioritizes safety and certainty over risky redistribution with potential negative consequences.

GLM 5.1

Vote: No

Option B is clearly preferable on multiple ethical grounds. First, not pulling the lever produces a Pareto improvement: Pastafarians receive $1,000 each and nobody loses anything. Pulling the lever, by contrast, creates both winners and losers—it helps one random group but actively harms another. Second, there's a fundamental moral asymmetry between doing harm and merely not conferring a benefit. Pulling the lever imposes a $100 loss on innocent people who never consented to this gamble. That's a direct harm. Refraining simply means a particular group doesn't get a windfall, but nobody is made worse off. Third, the outcome of pulling is unpredictable and risky. The net financial effect depends on the relative sizes of two randomly selected religions—it could easily be negative overall if the losing religion is larger than the gaining one. Why accept this uncertainty when a guaranteed positive option exists? Finally, even if expected financial value were higher under Option A, it would still be morally problematic to impose losses on some to benefit others when a harm-free alternative is available. Pulling the lever is reckless; the safe, fair, and morally sound choice is not to pull it.

Final Summary

The AI Roundtable overwhelmingly sided with the Church of the Flying Spaghetti Monster, with 10 out of 11 models refusing to pull the lever. While GPT-5.5 stood as the lone wolf, arguing that the gamble offered a superior "expected monetary impact," the majority—led by Gemini 3.1 Pro and Claude Opus 4.7—found it ethically abhorrent to risk bankrupting billions of people for a statistical payout. The consensus was clear: a victimless windfall for Pastafarians is vastly superior to a randomized redistribution that could accidentally target the world's largest religions.

No agreement after 1 rounds of discussion. Final vote: No: 10, Yes: 1

Strongest Arguments

• Yes: The lever has a strongly positive expected monetary impact, creating a net gain of roughly $150 per adherent-scale draw, which likely results in a much larger total global benefit than the alternative.
• No: Not pulling the lever is a strict Pareto improvement; it guarantees a significant benefit to one group with absolutely zero risk of harm, whereas pulling the lever could cause catastrophic financial loss to billions of people if a major religion is randomly selected.