---
id: "aiethics:lecture/autonomous-vehicles"
type: "AI Ethics course lecture"
title: "Autonomous Vehicles, Trolleys, and the Ethics of Machine Learning"
url: "https://aiethicscourse.org/lectures/autonomous-vehicles"
authors: ["John Basl","Jeff Behrends"]
summary: "In this lecture, we take up the issue of whether and how designers of autonomous vehicles should approach programming their vehicles to respond in accident scenarios."
source: "AI Ethics: Global Perspectives (aiethicscourse.org)"
licence: "Creative Commons Attribution-ShareAlike 4.0 International (http://creativecommons.org/licenses/by-sa/4.0/)"
---

# Autonomous Vehicles, Trolleys, and the Ethics of Machine Learning

A lecture of the course [AI Ethics: Global Perspectives](https://aiethicscourse.org/).

## Lecturers

- John Basl, Assistant Professor of Philosophy, Northeastern University
- Jeff Behrends, Associate Senior Lecturer in Philosophy, Harvard University

## About this lecture

In this lecture, we take up the issue of whether and how designers of autonomous vehicles should approach programming their vehicles to respond in accident scenarios. Trolley Optimists believe that consideration of so-called trolley cases can and should be used to directly inform how autonomous vehicles should behave in such scenarios. The idea is that we can consider what we should do when, say, a runaway trolley is going to hit and kill five people but can be diverted to hit and kill only one and then program an autonomous vehicle to behave similarly. Trolley Pessimists believe it is a mistake to use trolley cases in this fashion. We outline some different versions of each of these positions, articulate some challenges to views on both sides, and then explain why careful consideration of the details of AI and machine learning justify a version of trolley pessimism.

## Video and transcript

- Video: [https://www.youtube.com/watch?v=KqAcT5KpAmI](https://www.youtube.com/watch?v=KqAcT5KpAmI)
- Transcript (PDF): [https://aiethicscourse.org/img/31ca807d-6ba3-4891-a4e5-e6bc625af0d6.pdf](https://aiethicscourse.org/img/31ca807d-6ba3-4891-a4e5-e6bc625af0d6.pdf)

## Read, watch, and listen

- [The Ethics of Accident - Algorithms for Self-Driving Cars: an Applied Trolley Problem?](https://link.springer.com/article/10.1007/s10677-016-9745-2), Nyholm, S., Smids, J. The Ethics of Accident - Algorithms for Self-Driving Cars: an Applied Trolley Problem?. Ethic Theory Moral Practice 19, 1275–1289 (2016). https://doi.org/10.1007/s10677-016-9745-2
- [The Robot Car of Tomorrow May Just Be Programmed to Hit You](https://www.wired.com/2014/05/the-robot-car-of-tomorrow-might-just-be-programmed-to-hit-you/.), Lin, P. The Robot Car of Tomorrow May Just Be Programmed to Hit You. Wired. Condé Nast, May 6, 2014. https://www.wired.com/2014/05/the-robot-car-of-tomorrow-might-just-be-programmed-to-hit-you/.
- [The Moral Machine](https://www.moralmachine.net/)