# Should Self-Driving Cars Be Programmed to Sacrifice Passengers to Save Pedestrians?

> Analyze the classic 'trolley problem' of autonomous vehicles: should AV algorithms prioritize protecting the occupant or minimizing total casualties?

- Category: Artificial Intelligence
- Difficulty: hard
- Suitable for: High School, College, Adults
- Web page: https://debatetopicideas.com/debate-topics/self-driving-cars-moral-programming/

## Arguments for

- **Utilitarian ethics dictates minimizing total human loss of life**: If an autonomous vehicle must choose between swerving into a tree killing one passenger or plowing into a crowd of five schoolchildren, minimizing deaths is correct.
- **Pedestrians are completely innocent and vulnerable road users**: Pedestrians did not choose to enter the heavy motor vehicle and lack reinforced crumple zones, airbags, and seatbelts that protect car occupants.
- **Society will reject autonomous driving without clear ethical rules**: Public trust demands that automated software adhere to transparent, democratically codified moral guidelines rather than corporate secretive code.
- **Predictable algorithmic responses reduce chaotic secondary crashes**: Carefully modeled emergency swerves calculate friction and surrounding traffic far better than panic-stricken human driver reactions.

## Arguments against

- **No consumer will purchase a vehicle programmed to kill them**: Surveys prove people overwhelmingly support utilitarian cars in the abstract, but refuse to buy any car that would sacrifice their own children in an emergency.
- **Vehicle manufacturers owe a fiduciary duty of protection to customers**: Automakers enter a contract with the vehicle buyer; programming a consumer product to intentionally harm its owner creates fatal product liability.
- **Real-world crash physics rarely mirror binary philosophical thought experiments**: Engineers focus on braking, traction control, and collision avoidance; algorithmic 'moral triage' assumes impossible mathematical clairvoyance.
- **Slowing adoption of self-driving cars costs tens of thousands of lives**: Human drivers kill 40,000 people annually through drunk, distracted, and reckless driving; paralyzing AV rollouts over rare trolley dilemmas causes mass deaths.

## Counter-questions for cross-examination

- If a self-driving car must crash into a motorcyclist with a helmet vs one without, which should it choose?
- Should governments allow buyers to adjust an 'ethics slider' between personal protection and public utilitarian safety?
- Why did Germany's ethics commission rule that autonomous vehicles must never factor in a victim's age, gender, or social status?
- Who is legally liable if a vehicle makes an algorithmic decision that sacrifices an occupant: the engineer or the passenger?
- Are hypothetical trolley problems a useful engineering framework or an academic distraction from building basic collision-avoidance braking?

## Research sources

- <https://www.nature.com/articles/s41586-018-0637-6>
- <https://www.media.mit.edu/projects/moral-machine/overview/>

## Related debate topics

- [Should Governments Strictly Regulate Artificial Intelligence?](https://debatetopicideas.com/debate-topics/artificial-intelligence-regulation.md)
- [Should Governments Ban Facial Recognition Technology in Public Spaces?](https://debatetopicideas.com/debate-topics/facial-recognition-public-surveillance-ban.md)
- [Will Advanced Automation and AI Permanently Reduce Total Human Employment?](https://debatetopicideas.com/debate-topics/automation-replacing-human-labor.md)
- [Should Fully Autonomous Lethal Weapons Be Banned by International Treaty?](https://debatetopicideas.com/debate-topics/autonomous-weapons-ban.md)
- [Should Creating Non-Consensual Deepfakes Be Classified as a Felony Crime?](https://debatetopicideas.com/debate-topics/deepfakes-criminalization-laws.md)

---

Source: Debate Topics (https://debatetopicideas.com/). Free to use for classroom and tournament practice.
