When Fatal Crashes Can't Be Avoided, Who Should Self-Driving Cars Save? Or Sacrifice? Results From A Global Survey May Surprise You
Monday, November 05, 2018
Experts predict that there will be 10 million self-driving cars on the roads by 2020. Any outstanding issues need to be resolved before then. One outstanding issue is the "trolley problem" - a situation where a fatal vehicle crash can not be avoided and the self-driving car must decide whether to save the passenger or a nearby pedestrian. Ethical issues with self-driving cars are not new. There are related issues, and some experts have called for a code of ethics.
Like it or not, the software in self-driving cars must be programmed to make decisions like this. Which person in a "trolley problem" should the self-driving car save? In other words, the software must be programmed with moral preferences which dictate which person to sacrifice.
The answer is tricky. You might assume: always save the driver, since nobody would buy self-driving car which would kill their owners. What if the pedestrian is crossing against a 'do not cross' signal within a crosswalk? Does the answer change if there are multiple pedestrians in the crosswalk? What if the pedestrians are children, elders, or pregnant? Or a doctor? Does it matter if the passenger is older than the pedestrians?
To understand what the public wants -- expects -- in self-driving cars, also known as autonomous vehicles (AV), researchers from MIT asked consumers in a massive, online global survey. The survey included 2 million people from 233 countries. The survey included 13 accident scenarios with nine varying factors:
- "Sparing people versus pets/animals,
- Staying on course versus swerving,
- Sparing passengers versus pedestrians,
- Sparing more lives versus fewer lives,
- Sparing men versus women,
- Sparing the young versus the elderly,
- Sparing pedestrians who cross legally versus jaywalking,
- Sparing the fit versus the less fit, and
- Sparing those with higher social status versus lower social status."
Besides recording the accident choices, the researchers also collected demographic information (e.g., gender, age, income, education, attitudes about religion and politics, geo-location) about the survey participants, in order to identify clusters: groups, areas, countries, territories, or regions containing people with similar "moral preferences."
"The study is basically trying to understand the kinds of moral decisions that driverless cars might have to resort to," Edmond Awad, lead author of the study from the MIT Media Lab, said in a statement. "We don't know yet how they should do that."
And the overall findings:
"First, human lives should be spared over those of animals; many people should be saved over a few; and younger people should be preserved ahead of the elderly."
These have implications for policymakers. The researchers noted:
"... given the strong preference for sparing children, policymakers must be aware of a dual challenge if they decide not to give a special status to children: the challenge of explaining the rationale for such a decision, and the challenge of handling the strong backlash that will inevitably occur the day an autonomous vehicle sacrifices children in a dilemma situation."
The researchers found regional differences about who should be saved:
"The first cluster (which we label the Western cluster) contains North America as well as many European countries of Protestant, Catholic, and Orthodox Christian cultural groups. The internal structure within this cluster also exhibits notable face validity, with a sub-cluster containing Scandinavian countries, and a sub-cluster containing Commonwealth countries.
The second cluster (which we call the Eastern cluster) contains many far eastern countries such as Japan and Taiwan that belong to the Confucianist cultural group, and Islamic countries such as Indonesia, Pakistan and Saudi Arabia.
The third cluster (a broadly Southern cluster) consists of the Latin American countries of Central and South America, in addition to some countries that are characterized in part by French influence (for example, metropolitan France, French overseas territories, and territories that were at some point under French leadership). Latin American countries are cleanly separated in their own sub-cluster within the Southern cluster."
The researchers also observed:
"... systematic differences between individualistic cultures and collectivistic cultures. Participants from individualistic cultures, which emphasize the distinctive value of each individual, show a stronger preference for sparing the greater number of characters. Furthermore, participants from collectivistic cultures, which emphasize the respect that is due to older members of the community, show a weaker preference for sparing younger characters... prosperity (as indexed by GDP per capita) and the quality of rules and institutions (as indexed by the Rule of Law) correlate with a greater preference against pedestrians who cross illegally. In other words, participants from countries that are poorer and suffer from weaker institutions are more tolerant of pedestrians who cross illegally, presumably because of their experience of lower rule compliance and weaker punishment of rule deviation... higher country-level economic inequality (as indexed by the country’s Gini coefficient) corresponds to how unequally characters of different social status are treated. Those from countries with less economic equality between the rich and poor also treat the rich and poor less equally... In nearly all countries, participants showed a preference for female characters; however, this preference was stronger in nations with better health and survival prospects for women. In other words, in places where there is less devaluation of women’s lives in health and at birth, males are seen as more expendable..."
This is huge. It makes one question the wisdom of a one-size-fits-all programming approach by AV makers wishing to sell cars globally. Citizens in clusters may resent an AV maker forcing its moral preferences upon them. Some clusters or countries may demand vehicles matching their moral preferences.
The researchers concluded (emphasis added):
"Never in the history of humanity have we allowed a machine to autonomously decide who should live and who should die, in a fraction of a second, without real-time supervision. We are going to cross that bridge any time now, and it will not happen in a distant theatre of military operations; it will happen in that most mundane aspect of our lives, everyday transportation. Before we allow our cars to make ethical decisions, we need to have a global conversation to express our preferences to the companies that will design moral algorithms, and to the policymakers that will regulate them... Our data helped us to identify three strong preferences that can serve as building blocks for discussions of universal machine ethics, even if they are not ultimately endorsed by policymakers: the preference for sparing human lives, the preference for sparing more lives, and the preference for sparing young lives. Some preferences based on gender or social status vary considerably across countries, and appear to reflect underlying societal-level preferences..."
And the researchers advised caution, given this study's limitations (emphasis added):
"Even with a sample size as large as ours, we could not do justice to all of the complexity of autonomous vehicle dilemmas. For example, we did not introduce uncertainty about the fates of the characters, and we did not introduce any uncertainty about the classification of these characters. In our scenarios, characters were recognized as adults, children, and so on with 100% certainty, and life-and-death outcomes were predicted with 100% certainty. These assumptions are technologically unrealistic, but they were necessary... Similarly, we did not manipulate the hypothetical relationship between respondents and characters (for example, relatives or spouses)... Indeed, we can embrace the challenges of machine ethics as a unique opportunity to decide, as a community, what we believe to be right or wrong; and to make sure that machines, unlike humans, unerringly follow these moral preferences. We might not reach universal agreement: even the strongest preferences expressed through the [survey] showed substantial cultural variations..."
Several important limitations to remember. And, there are more. It didn't address self-driving trucks. Should an AV tractor-trailer semi -- often called a robotruck -- carrying $2 million worth of goods sacrifice its load (and passenger) to save one or more pedestrians? What about one or more drivers on the highway? Does it matter if the other drivers are motorcyclists, school buses, or ambulances?
What about autonomous freighters? Should an AV cargo ship be programed to sacrifice its $80 million load to save a pleasure craft? Does the size (e.g., number of passengers) of the pleasure craft matter? What if the other craft is a cabin cruiser with five persons? Or a cruise ship with 2,000 passengers and a crew of 800? What happens in international waters between AV ships from different countries programmed with different moral preferences?
Regardless, this MIT research seems invaluable. It's a good start. AV makers (e.g., autos, ships, trucks) need to explicitly state what their vehicles will (and won't do). Don't hide behind legalese similar to what exists today in too many online terms-of-use and privacy policies.
Hopefully, corporate executives and government policymakers will listen, consider the limitations, demand follow-up research, and not dive headlong into the AV pool without looking first. After reading this study, it struck me that similar research would have been wise before building a global social media service, since people in different countries or regions having varying preferences with online privacy, sharing information, and corporate surveillance. What are your opinions?
You can follow this conversation by subscribing to the comment feed for this post.