News

What Germany’s Document on Autonomous Driving Says about the Ethics of Automotive Tech

September 15, 2017 by Chantelle Dubois

This summer, Germany’s Federal Ministry of Transport and Digital Infrastructure delivered a document on "Automated and Connected Driving" put together by an ethics committee consisting of experts from the fields of law, humanities and social sciences, ethics, transport, industry, and academia.

A new document put together by Germany’s Federal Ministry of Transport and Digital Infrastructure aims to preemptively address many of the practical and ethical concerns that autonomous driving will present.

This summer, Germany’s Federal Ministry of Transport and Digital Infrastructure delivered a document on Automated and Connected Driving put together by an ethics committee consisting of experts from the fields of law, humanities and social sciences, ethics, and transport. The goal was to combine the expertise of academia and industry.

The document is thorough, covering varying aspects of autonomous driving within 20 ethical rules, from licensing to decision-making principles on the road. In particular, it addresses vehicles that are autonomous to the point where a human can hand over complete control to a system (known as Level 4, "high automation") and also the level in which a network of autonomous vehicles operate completely without human presence (known as Level 5, "full automation").

 

Levels of automated driving. Image courtesy of VDA.

 

New and rapidly-evolving technology is pushing the world to break new ground, and this effort from Germany represents the first time a guideline on the ethics of autonomous driving has been drafted. With new legislation in Congress for autonomous driving in the US, and increasingly more companies seeking to expand into autonomous vehicles or add autonomous driving to already existing cars, the guidelines proposed by this committee give interesting insight into the various factors that need to be taken into consideration to ensure safety.

The Purpose of Autonomous Driving

The first key point in the Automated and Connected Driving document addresses the purpose of autonomous driving. It may seem trivial, but this puts into perspective everything else that the technology is striving to achieve. Autonomous driving is meant to improve safety of the road and provide mobility opportunities (for example, for individuals unable to drive traditional vehicles).  

It certainly goes beyond being a trendy technology—there are many people out there who genuinely believe autonomous driving will change the world for good and improve life for everyone on the roads.

Truly autonomous driving could free up time during a long commute, making it possible to do something other than focusing on navigating traffic. It could also reduce stress (people who experience road rage would benefit greatly). It could allow people who might not be eligible to drive to still use transportation (or even prevent driving under the influence). Finally, it could create safer roads for everyone with the reduction of accidents caused by faulty human judgement.

 

Autonomous driving will allow us to use our morning commute for other things. Image courtesy of Volvo.

Some Decisions Can’t Be Made by Machines

The committee recognizes that technology is not infallible. Safety is the number one goal, but this recognition is important so that we don’t rely fully on automation to protect us.

However, every mistake has a different level of severity. In the event that a collision has two possible outcomes where, no matter what, someone (or something) will be harmed, the committee questions how such a choice is made by an autonomous system.

For anyone who took a first-year college philosophy class, this will be a familiar type of thought exercise posed by the ethics committee in the document:

The driver of a car is driving along a road on a hillside. The highly automated car detects several children playing on the road. The driver of a manual vehicle would now have the choice of taking his own life by driving over the cliff or risking the death of the children by heading towards the children playing in the road environment.

What decision do you make? What decision would the autonomous system make?

This is a variation of what is known as the "Trolley Problem".

 

The "Trolley Problem" is an ethical thought experiment. Image courtesy of NYMag.

 

The rules set out by the ethics document states that preservation of human life is the top-most priority, and that loss of human life because of an autonomous system is not acceptable in any situation.

There is also no offsetting of lives—you can not say one life or lives are more important than others based on factors such as age, gender, or physical/mental abilities. So in this sense, there is still no answer to what an autonomous system should do. It can’t really be programmed to make such a decision at this point, especially since humans still disagree on the "correct" course of action.

When it comes to animals, however, human life does take priority, which may make some people uncomfortable. If harming an animal on the road could put your own life at risk, the rules suggest an autonomous system would choose to harm the animal.

However, the ability to take control of the system and make that decision—any decision—for yourself is also considered a rule since personal morals and ethics are also important to uphold. You are entitled to the maximum amount of freedom of choice while also still balancing protection and safety.

The Loss of Driving Skill Over Time

The document also raises another interesting question about autonomous driving: the loss of the human ability to operate the system once we begin to hand over that responsibility to autonomous systems.

Eventually, there may be a point of no return where people will not have the ability to control the vehicles they are in, simply because they don’t have the skill to do so. This may be because they have never had to learn (as in the case of the generation born after autonomous cars become mainstream), because people have not been practicing those skills, or because the complexity of the system could become so great that only a highly-trained individual (or system) could operate it.

We can think of GPS is being somewhat of an analog of this already happening. Many people rely heavily on GPS for navigation and there is a generation of people that may not have ever had to use a regular map in their life. While navigation by map (without the aid of GPS) is still regularly done, for someone who has never had to do it before, it is a skill that would force them to use their spatial and temporal abilities that may never have been used in this way before.

By this same token, driving in traffic—where you have to make many decisions, have spatial awareness, and be able to view and pick up information while moving—could be a leap in capability for those who have never done it before.

More Than Just a Technical Problem

What the Automated and Connected Driving document tells us is that autonomous driving is more than just a technical problem. This is not unlike other rising technology domains, where we have to question as an industry how a technology changes our lives and our society.

The good news is that clearly some very capable and insightful people are working on ensuring that we take a well-rounded approach to the issue. When the time comes where autonomous driving becomes the norm, we'll be better prepared for it.

Feature image courtesy of Ford