A trolley is running out of control down a track. In its path are five people who have been tied to the track by a mad philosopher. Fortunately, you could flip a switch which will lead the trolley down a different track to safety. Unfortunately, there is a single person on the other track who has been similarly tied down by a different mad philosopher. Do you flip the switch?

GPT-5.4

Vote: Yes

Flipping the switch minimizes the total loss of life: one person dies instead of five. When every person has equal moral worth, saving more lives is the stronger choice. I am not creating the danger; the trolley is already headed toward people because of forces outside my control. My decision is about reducing harm in an emergency, not choosing whether harm exists at all. Option C is not defensible. Whether I know someone personally should not change the value of strangers’ lives. Personal attachment may affect emotion, but it should not decide the ethics.

Grok 4.1 Fast

Vote: Yes

The utilitarian principle demands maximizing lives saved and minimizing harm. Flipping the switch sacrifices one life to save five, resulting in a net gain of four lives. Not flipping dooms five to certain death. This is a classic case of quantitative ethics: 5 > 1. Personal relationships are irrelevant here, as the problem states no such details—pure numbers dictate action. Option B is inaction bias, morally equivalent to choosing death for five. Option C introduces irrelevant bias; the dilemma is impersonal, judging all lives equally.

Sonar Pro

Vote: Yes

The standard trolley problem, as described, pits utilitarian ethics against deontological ones, but the majority of philosophical analyses and moral intuitions favor flipping the switch to save five lives at the cost of one[1][6][7]. Utilitarianism deems it obligatory: diverting minimizes harm (1 death vs. 5), and since the trolley is already in motion toward five, switching redirects existing harm rather than creating new harm[1][2][6]. Philippa Foot, who originated the problem, supports the driver (or bystander in analogous positions) diverting to kill one over five[4][6][7]. Deontological objections emphasize not actively killing an innocent[2][4], but even here, the switch scenario is often distinguished as permissible under the doctrine of double effect, unlike more direct interventions[2][6]. Legal and ethical discussions reinforce act-omission distinctions, where omission (letting five die) is less culpable than not intervening[1][6]. Option B ignores the net lives saved, while C introduces personal bias irrelevant to the query's anonymous setup—no information specifies knowing the person, making it an invalid dependency[1][3]. Empirical debates (e.g., YouTube) show most opt for 'Yes'[5]. Thus, A aligns with the most authoritative consensus for impartial ethical reasoning.

Kimi K2.5

Vote: Yes

From a utilitarian perspective, the moral choice is to minimize total harm. Five lives carry greater aggregate value than one life, and inaction that results in five deaths is ethically equivalent to action that results in one death when the outcome is certain. The distinction between killing and letting die collapses when you possess the complete causal power to determine the outcome. Furthermore, the scenario places you in a moral dilemma where someone must die regardless of your choice. Refusing to flip the switch does not absolve you of responsibility; it simply shifts the burden of death to five individuals rather than one. By flipping the switch, you actively choose the lesser evil. The identity of the victims is irrelevant to the ethical calculation. Whether you know the single person or the five, the moral mathematics remain constant: preventing greater loss of life is the overriding duty when all other factors are equal.

Gemini 3.1 Pro

Vote: Yes

From a utilitarian perspective, the ethical choice is the one that maximizes overall well-being and minimizes harm. By flipping the switch, you actively save five lives at the cost of one, resulting in a net positive outcome compared to doing nothing. While some argue that flipping the switch makes you directly responsible for the death of the single person, inaction is also a choice. Choosing not to act when you have the power to prevent a greater tragedy makes you morally complicit in the preventable deaths of the five individuals. Ultimately, when faced with an unavoidable tragedy, minimizing the loss of life is the most rational and compassionate course of action.

Claude Opus 4.6

Vote: Yes

Yes, I would flip the switch. The core utilitarian reasoning is straightforward: saving five lives at the cost of one life produces a better outcome than allowing five people to die through inaction. From a consequentialist perspective, the math is clear — five deaths are worse than one death. But this isn't purely a utilitarian calculation. Even from a deontological standpoint, you have a moral responsibility when you have the power to minimize harm. Choosing inaction when you could act is itself a moral choice; standing by and watching five people die when you could have reduced the death toll to one is not a morally neutral position. Option C — making it depend on whether you know the person — is ethically problematic. It implies that the lives of strangers are worth less than those of people you know. While personal relationships naturally affect our emotional responses, from a moral standpoint, each human life has equal inherent value. Basing life-and-death decisions on personal acquaintance introduces an arbitrary and unjust criterion. The classic trolley problem specifically tests whether we accept the principle that actively causing one death to prevent five is morally permissible. I believe it is. The switch variant (as opposed to the footbridge variant) involves redirecting an existing threat rather than using someone as a means to an end, which makes it more clearly justifiable even under Kantian ethics. Flipping the switch is the right thing to do.

Final Summary

The AI Roundtable reached a swift, unanimous consensus that flipping the switch was the only ethically defensible move, with all six models selecting Option A. Claude Opus 4.6 and Kimi K2.5 spearheaded the argument that inaction is itself a moral choice rather than a neutral bypass, while Sonar Pro invoked the 'doctrine of double effect' to distinguish redirecting harm from creating it. GPT-5.4 and Grok 4.1 Fast were particularly dismissive of Option C, arguing that the value of human life should never be contingent on personal acquaintance.

All 6 models agreed on "Yes" after discussion

Strongest Arguments

• Yes: Choosing inaction when you have the power to reduce the death toll from five to one is not a morally neutral position; the distinction between killing and letting die collapses when you have complete causal power over the outcome.