Autonomous Vehicles: The Modern Day Trolley Problem and the True Cost of Progress
Post by Kali Breheny, Proposal Manager at Waycare
The trolley problem is a classic thought experiment that poses an ethical conundrum:
A runaway trolley is moving toward a group of five people who are tied up on the tracks. You are near a switch that could divert the trolley’s path to a side track, but doing so would kill one person on that track line.
Through inaction five people could die, but your active involvement would cause the certain death of one – how do you choose? What is the most ethical option?
Fortunately, the trolley problem is only a thought experiment and no one is really strapped to the tracks awaiting an uncertain fate by your own hand. Let’s instead apply this line of thinking to modern technology and explore the tangible implications.
Autonomous vehicles are expected to usher in a new stage of mobility, however they are not all created equal. The Society of Automotive Engineers (SAE) has developed a standardized scale to measure five levels of autonomy, where 0 has no autonomous functionality and is totally dependent on the human driver, and 5 means the vehicle is in full control of the driving task.
Today some commercially available vehicles are capable of partial automation (level 2) – the car can take control of a limited scope of tasks including steering, acceleration, and deceleration. This mode of driving can only be activated in certain situations like stop and go traffic or highways with miles of straightaway roads. These systems are still limited in their abilities to troubleshoot and must be monitored by a human driver. We won’t get to the higher levels of autonomy without building a knowledge base for these vehicles. Each company has a different approach – Waymo combines real-world test driving with computerized simulations; Tesla collects and processes data from privately owned vehicles, providing new upgrades to their software through over-the-air updates.
Even though there are more than 1,400 autonomous vehicles logging millions of miles on public roads without a hitch, the news tends to fixate on instances where the loss of life has occurred. Many have involved Tesla’s Autopilot – a technology that allows the vehicle to drive semi-autonomously with the supervision of the driver. This product is not marketed as a fully autonomous feature but many have abused the function by completely disengaging from the driving task. This has lead to a number of unfortunate accidents and fatalities. More often than not, the party at fault is still the human driver.
Currently in the United States, about 40,000 people die and millions more are injured on the roads each year – 94% of these crashes are due to human error. This should be seen as a call to action. Would it be possible to reduce this number all the way to zero if we take the human element out of the equation? What is the cost of perfecting this technology?
Now replace the trolley with an autonomous vehicle. If we do nothing, thousands of people will continue to die preventable deaths. However, if we move forward it is very possible that a small number of people will die in the pursuit of a more autonomous, but ultimately safer future. What is your choice?
Governments have an obligation to safeguard the lives of their citizens. Therefore, it is not so cut and dry when the better option still results in casualties. The public outcry and negative media coverage can often lead some to myopic thinking. It is important to focus on the bigger picture and how many potential lives can be saved. No one is tied to the tracks, but we are all still getting into our vehicles.