Wait, what?
You heard me right: cats and dogs are relatively safer than humans while crossing roads in India.
How did you arrive at that determination?
That’s the conclusion derived from a research spearheaded by the Massachusetts Institute of Technology, whose findings were published last fortnight in Nature magazine. The study was centred around the ethical issues that need to be addressed while configuring “autonomous” (or self-driven) vehicles.
What do cars on autopilot have to do with ethics?
Plenty. Governments around the world are acknowledging that mobility is increasingly being shaped by the digital revolution. Particularly given the imminent licensing of self-driven automobiles, it raises moral questions about the factors that will underlie the decisions that an Artificial Intelligence program (which drives the car) makes in the interest of enhanced road safety.
I still don’t get it. What’s the moral issue?
Let me put it bluntly. Say you’re driving a car and your brakes fail, and you find the car is speeding towards a throng of five elderly people. If you could steer the car away and save them, you would of course. But if in making that choice, you are forced to drive into a group of three children standing by the roadside, would you be conflicted? Alternatively, what if in trying to avoid hitting a child, you run into a group of pets? Is that an acceptable trade-off? What if you had to choose between running over a well-dressed business executive jumping a pedestrian red light — and a homeless person? You have to decide who lives and who dies.
Hmm, I see where this is going.
The AI program that drives cars has to be programmed to make decisions like these for various permutations of people and situations. It’s a variant of the ‘Trolley problem’.
What’s that?
It’s a thought experiment used in moral psychology. If you see a runaway trolley and have to choose between doing nothing (thereby seeing five people being killed) or redirecting the trolley onto a side track, where a single person will be killed, what would be the more “ethical” option?
Okay, so what’s this about cats and children in India?
For their research, the scientists surveyed some 40 million volunteers online using their ‘Moral Machine’ program; people from 233 countries or territories answered a set of questions on various scenarios involving a runaway driverless car and different groups of people who would be killed as a consequence. Their answers showed that there are geography-specific cultural values underlying the ethical decisions that people make in such situations. Given the option between sparing humans or pets from a runaway car, Indians afforded a lower priority to humans. Youngsters ranked even lower in their list of priority for saving, as compared to the elderly.
Ergo, cats are safer than children on Indian roads?
Yes. The researchers found three broad patterns. Across the world, those who were surveyed opted to save humans (over animals), save a bigger number of people when given an option, and save children (over elderly people). But countries in South Asia and East Asia, where elders are held in respect, do not have a strong preference for saving youngsters’ lives.
And in countries with higher inequality, more people were willing to save the executive at the cost of the homeless. The needle of Indians’ ethical compass too points in that direction.
Can I too take the survey?
Sure, on moralmachine.mit.edu; go right ahead and play God with other people’s lives.
A weekly column that helps you ask the right questions