Ethical Decision Policy for Self-Driving Cars When Every Option Causes Harm
Company: Google
Role: Product Manager
Category: Product / Decision Making
Difficulty: medium
Interview Round: Other
You are advising a company that builds autonomous vehicles on the ethics of its driving software. Some situations leave the vehicle with no harmless option: for example, swerving to avoid a pedestrian may put the people inside the car at risk. How would you handle situations like this?
Explain the principles you would want the vehicle to follow, how those principles become concrete product and engineering decisions, who should get to decide, and how you would explain and defend the approach to riders, the public and regulators.
```hint Look beyond the split second
Ask how often a true either-or choice actually arises, what the vehicle can do earlier to make it rarer, and how much the car can really know about the outcome of each option in the moment.
```
### Clarifying Questions
- Is the vehicle a driverless ride-hailing car carrying passengers, or a privately owned car whose owner may expect to be protected first?
- Is the program still testing with safety drivers on board, or already carrying riders with no one behind the wheel?
- Which countries or states will the vehicles operate in, and do their laws or regulators already constrain how such collisions must be handled?
- Is my job to set the ethical policy itself, or to design the process through which the company and outside parties set it?
### What a Strong Answer Covers
- Reframing rare forced-choice dilemmas as one part of the vehicle's overall risk policy, including how prevention makes them rarer
- A small set of explicit principles for unavoidable-harm cases and a stated way to resolve conflicts among them, reasoned under uncertain perception
- How the principles become testable requirements, simulation scenarios and release criteria
- Governance and accountability: who sets and approves the policy, the role of regulators, incident review, and transparency to riders and the public
- Evidence and metrics showing the policy works, and the trust, adoption and liability trade-offs it creates
### Follow-up Questions
- Should buyers be allowed to set their car to prioritize its occupants? Defend your answer.
- After a crash in which the car swerved and a passenger was hurt, how would you investigate, and what would you say publicly?
- Should the vehicle behave the same way in every country if surveys show moral preferences differ across cultures?
Overview: Product manager interview question on how a self-driving car should act when every option carries harm, such as swerving to spare a pedestrian while risking its passengers. It tests ethical reasoning under uncertainty, turning principles into engineering requirements, governance, and public trust.