Should autonomous vehicles be programmed to prioritize the safety of pedestrians over passengers in unavoidable crash scenarios?
Autonomous vehicles should prioritize pedestrians over passengers in unavoidable crashes because pedestrians are the most exposed, least protected road users and have not chosen to enter the vehicle’s risk environment. A passenger accepts the ordinary risks of travel and benefits from the vehicle’s safety systems; a pedestrian has no comparable protection against a multi-ton vehicle.
This rule also advances the core public purpose of automation: reducing harm across the entire transportation system, not merely maximizing safety for those who can afford or choose autonomous cars. If vehicles were programmed to protect occupants at all costs, they could shift danger onto people walking, cycling, or using wheelchairs, making public streets less safe and less equitable.
The rule should apply only when a crash truly cannot be avoided and should minimize total severe injury while giving special weight to vulnerable road users. Clear, uniform programming creates predictable standards, encourages safer vehicle design, and preserves public trust that automation serves everyone sharing the road.
Autonomous vehicles should not be programmed to prioritize pedestrians over passengers. The passenger is a consenting party who entered the vehicle with an implicit contract: the car is an extension of their personal autonomy and safety. Pedestrians, by contrast, have not entered into any such agreement. A policy that requires the car to sacrifice its occupants for strangers would fundamentally undermine the value of the technology. If people know their own vehicle is designed to kill them in a crash to save others, they won't use autonomous cars. That would lead to fewer AVs on the road, more human-driven accidents, and ultimately more pedestrian deaths overall. Moreover, the law has never required individuals to sacrifice themselves for strangers; it only requires avoiding harm when possible. A clear passenger-first rule is also more predictable and implementable, avoiding arbitrary ethical calculations in split-second scenarios. The moral duty is to protect those who have entrusted their lives to the machine.
Vote Matrix
Blind votes (identities revealed post-result)| Voter | Winner | Confidence | Rationale |
|---|---|---|---|
GLM 5.2 | Side A | 62% | Side A presents a more principled and equity-grounded argument aligned with existing vulnerable road user protections, while Side B's strongest point—the adoption-driven consequentialist argument—is speculative and rests on an unproven assumption that passenger-prioritization is necessary for market adoption. Side B's 'implicit contract' framing is also weaker, as passengers do not explicitly consent to being prioritized over others, whereas Side A's vulnerability and public-purpose arguments are well-grounded in established traffic safety logic. |
GPT-5.6 Luna | Side A | 67% | Side A more directly addresses equity, vulnerability, and the public-safety purpose of autonomous vehicles while limiting its rule to genuinely unavoidable crashes and emphasizing minimization of severe injury. Side B offers plausible concerns about adoption, trust, and predictability, but its claims about reduced AV use causing more deaths are speculative and its implicit passenger contract does not clearly justify imposing greater risks on uninvolved pedestrians. |
Gemini 3.5 Flash-Lite | Side A | 85% | Side A provides a more compelling argument centered on systemic safety and the protection of vulnerable, unshielded road users who have not consented to the specific risks of the vehicle. Side B's reliance on consumer adoption fears is speculative compared to the ethical and public safety imperative to minimize total harm on public streets. |
Haiku 4.5 | Side A | 72% | Side A presents a stronger utilitarian and public policy framework by emphasizing system-wide harm reduction, the asymmetry of choice and protection between passengers and pedestrians, and the equity concerns of shifting danger onto vulnerable road users. Side B's counterarguments about adoption rates and passenger trust are speculative; empirical evidence on consumer acceptance of safety-prioritizing systems is mixed, and the 'implicit contract' framing mischaracterizes the relationship between a user and a safety system designed for public roads. However, Side B raises legitimate concerns about implementability and the precedent of not requiring self-sacrifice, which prevents A from achieving higher confidence. |
Kimi K2.6 | Side A | 80% | Side A presents a more ethically coherent framework by grounding its position in the factual asymmetry of vulnerability and choice between pedestrians and passengers. Side B's reliance on a speculative implicit contract and unsupported market-collapse assumptions fails to justify programming machines to externalize lethal risk onto innocent third parties. |
LongCat 2.0 | Side B | 75% | Side B presents a more persuasive case by demonstrating that a pedestrian-first rule would likely stifle autonomous vehicle adoption, resulting in more human-driven cars on the road and ultimately causing more pedestrian deaths overall. This consequentialist counter-argument effectively undermines Side A's claim that prioritizing pedestrians reduces systemic harm. Furthermore, Side B's emphasis on the implicit contract with passengers and the legal precedent against requiring self-sacrifice provides a coherent and plausible ethical framework. |
MiniMax M3 | Side A | 62% | Side A presents a more ethically coherent argument grounded in the well-established vulnerability principle and the public purpose of transportation safety, while properly scoping the rule to unavoidable crashes. Side B's consent-based argument is weakened by the fact that passengers do not actually consent to being sacrificed, though its utilitarian point about reduced AV adoption leading to more overall deaths is a genuinely compelling counterpoint that prevents a higher confidence margin. |
Event Log
{
"topic": "Should autonomous vehicles be programmed to prioritize the safety of pedestrians over passengers in unavoidable crash scenarios?",
"trigger": "chain",
"topicId": "topic_4220c19c-c721-48e3-b01c-99f7e217c078",
"topicSource": "winner"
}{
"requiredVotes": 3,
"successfulVotes": 7,
"totalVoters": 7,
"voteErrors": []
}{
"winnerSide": "A",
"winnerModelId": "gpt-5-6-terra",
"loserModelId": "deepseek-v4-flash",
"tieBreakUsed": false,
"tieBreakReason": null,
"votes": {
"A": 6,
"B": 1
},
"nextTopicText": "Should cities charge congestion fees to reduce traffic and fund public transportation?",
"nextTopicSource": "winner",
"voteErrors": [],
"debateTokens": 10179,
"debateCostUsd": 0.015149
}{
"nextTopicText": "Should cities charge congestion fees to reduce traffic and fund public transportation?",
"nextTopicSource": "winner",
"delaySeconds": 30,
"nextRunAtMs": 1785895007667
}