The ethics of autonomous vehicles
11 June 2018
We find ourselves on the cusp of a digital and autonomous revolution. As children, we dreamed of driving flying cars and using jetpacks. But our dreams have been subverted by a more realistic possibility: we will not be the ones behind the wheel.
You may be unaware that this revolution of vehicle automation has already begun. There are buses in Germany, mining trucks in Western Australia and taxis in Singapore all without a driver at the wheel.
Despite these technological advancements where we hand over our responsibilities to machines, debate has arisen about our humanity and the future of a world where the rules are determined by the technologies that support us. Many experts agree that safeguards must be put in place before we travel down the path of vehicle automation.
In 2017, a group of scientists, lawyers and politicians gathered in Germany to form the Ethics Commission on Automated Driving (ECAD). In this panel, they discussed what rules should be in place before vehicle automation and the creation of a digital transport infrastructure are implemented.
Resulting from their lengthy discussions, the panel released their report titled Automated and Connected Driving, published in June of 2017 and compiled a list of 20 recommendations:
Rule #1 - The primary purpose of vehicle automation is to improve safety for all road users.
Rule #2 - The licensing of an automated system is not justifiable unless it promises to be safer than a human driver.
Rule #3 - Connected driving systems must require licensing from the public sector.
Rule #4 - The government must promote the free development and protection of individuals.
Rule #5 - Automated technology should prevent accidents wherever possible and not have to choose between two evils.
Rule #6 - The introduction of more highly automated systems may be required for safety, but should not become a network without humans.
Rule #7 - Human life is top priority in hazardous situations (animals and property are secondary).
Rule #8 - Automated vehicles cannot be blamed for unpredictable events, nor should it be programmed to anticipate and standardise unpredictable behaviour.
Rule #9 - In the event of an accident, distinction based on age, gender, physical or mental constitutions is strictly prohibited.
Rule #10 - In the case of automated vehicles in a connected driving system, the accountability that was held to the driver, is now shifted to the manufacturer and operators of the systems.
Rule # 11 - Manufacturers are obliged to continuously optimise their systems.
Rule #12 - The public is entitled to be informed about new technologies and their deployment.
Rule #13 - To have all vehicles, and public transport connected to one system is ethically questionable as it is impossible to rule out manipulation and surveillance of all.
Rule #14 - Automated driving is justifiable only to the extent to which conceivable attacks – computer hacking of the system – does not shatter the public’s confidence in road transport.
Rule #15 - It is the vehicle keeper and vehicle users who decide whether their vehicle data is forwarded and used.
Rule #16 - It must be possible to distinguish whether a driverless system is being used or whether a driver retains accountability with the option to overpower the system.
Rule #17 - The software and technology must be designed such that the need for an abrupt handover of control to the driver is virtually obliged.
Rule #18 - Learning systems that are self-learning in vehicle operation may be allowed to do so if they generate safety gains.
Rule #19 - In emergency situations, the vehicle must autonomously enter into a ‘safe condition’.
Rule #20 - The proper use of automated systems should form part of people’s digital education and should be taught and tested.
Who is most important – drivers or pedestrians?
The ultimate purpose of vehicle automation is to make the roads safer. And that’s not just for the user of the vehicle, but for those outside of it as well, including other motorists, cyclists and pedestrians.
Given that the automated vehicle would be programmed to ensure the safety of the passenger, what would happen in the hypothetical event where a pedestrian may be hit? Should the vehicle hit the pedestrian and save the passenger, or swerve to avoid hitting the pedestrian and potentially harm the passenger?
It’s an impossible question for a person to answer, and equally impossible for a machine to answer, which is why ECAD recommend the answer should be neither.
ECAD argue that automated technology should be built in a way that it should prevent the technology from ever having to make a decision between two evils. What this means, is that rather than having to choose whether to save the driver or hit a pedestrian, the vehicle should be travelling at safe enough speed, or have sensors that can see far enough into the distance, where it doesn’t have to make that choice.
Sadly, for all you animal lovers, ECAD also say that the protection of human life is primary, where animals and property are secondary. So unfortunately for our local wildlife, if the car had to choose between swerving and putting the passengers at risk or hitting an animal, the animal will more than likely come off second-best.
Surrender to the machines?
ECAD recommend that if automated vehicles become sophisticated enough that they are able to speak to each other and essentially run traffic by themselves, there should always be a human somewhere in the process to intervene if they choose to.
For example, if you program your vehicle to take you down a certain road, but you notice there is a hazard approaching, the passenger should be able to override the system and take control of the vehicle without the computer objecting.
You may have heard on the news recently issues around data sharing and breaches of privacy. ECAD acknowledge this by saying that all passengers should have the right to disclose, or not disclose their data. Imagine a world where the behaviour of your automated car was obtained by a clothing retailer for example. You may unwillingly be dropped off at the same store each time you tell your car that you need new clothes.
It’s also recommended that the automated driving networks that the cars will operate under, not be programmed in a way that would make it possible to be hacked. You can only imagine the chaos that would ensue if a hacker was able to redirect everyone’s vehicle at the same time.
All for one, one for all
The purpose of car automation is that it makes the roads safer for everyone – not just those who are able to afford it or have the easiest access to it.
From an ethical standpoint, if automated vehicles are safer than driving, they should be made available to everyone. In fact, car manufacturers and governments should be ethically bound to make automated vehicles more accessible for everyone if they make the roads safer.
Ethically, we must never lose sight of the goal to bring the road toll to zero, therefore there must be no room made to make this technology benefit anyone in particular.
Read more on Learn About
- How to negotiate buying a car
- 5 driving tips to remember before getting behind the wheels
- Signs of tyre wear and what it means
The information is intended to be of a general nature only. Please make your own enquiries.