Robotics And Ethics–The Smart Car

Google’s self-driving car will be roadworthy by the end of this year, and it is revolutionary.  We have known that the day would come when human drivers would be replaced, but perhaps not too soon.  The automated car will reduce accidents, improve efficiency, and reduce transportation costs.  Distracted driving will be a thing of the past, cars will be converted into dens, offices, and parlors, and a ride up the Jersey Turnpike will be a thing of beauty.


It is also no surprise that Google leads the ways in robotic cars, for big data is the key to artificial intelligence.  Not many years ago the chief researcher at Google, Peter Norvig, and the famed cognitive scientist, Noam Chomsky, duked it out at an MIT seminar about how to build a robot which could speak perfect, idiomatic, unaccented English, Chinese, or Kung! 

Chomsky said that one must understand the nature of human language, according to him an innate, pre-programmed ability linked to intellect, perception, logic, and versatility.  Norvig said, “Bah, humbug”. All that is required is a comprehensive study of all current English usage.  Record, categorize, and analyze the language of millions of speakers; and identify trends, idiomatic peculiarities, adherence to and diversion from standard grammar, accents, and context. Based on such big data and powerful analytical software, creating an artificially intelligent program that masters Ozark hillbilly or the Queen’s English is child’s play.



Google Earth is a popular app created by collecting trillions of bits of information by satellite or mobile camera.  Soon every street in America will have been surveyed, photographed, recorded, and posted electronically.


This particular technology is what makes smart cars possible. Before setting off on an automated journey, the car’s computer is loaded with the latest Google mapping data for its intended route with traffic lights, stop signs, crosswalks, shoulders, curves, and straightaways, and a normal traffic pattern.  As the car drives, its cameras view the scene in real time and compare it with the ideal.  A car entering the roundabout too fast; an edgy lane-changer; a wobbly beater in the passing lane.  The responsive software has been programmed with nearly limitless possibilities of caution – bad driving, potholes, rain, and poor lighting.  Because of the baseline data stored, it can focus on the aberrations.

So far, so good.  Everything in life can be disaggregated, categorized, and understood.  The fluent robot and the smart, safe car are based on the same principles – the collection of massive amounts of data, passed through algorithmic sieves, and recombined to create artificially intelligent machines.
The hardest part of artificial intelligence – at least in programs requiring reason and judgment, like driving a car – is how to choose between ethical alternatives. 



Does the smart car swerve to avoid hitting a woman with a baby carriage but upends a cyclist? A potential two deaths avoided are more valuable in ethical terms than one.  Breaking the law is considered unethical, but breaking a traffic law – such as jumping a divider – may be justified if it can avoid an accident and/or injury to passengers.

It might seem like a simple thought experiment, a twist on the classic “trolley problem,” an ethical conundrum that asks whether you’d save five people on a runaway trolley, at the price of killing one person on the tracks. But the more detailed the crash scenarios get, the harder they are to navigate. Assume that the robot has what can only be described as superhuman senses and reaction speed, thanks to its machine reflexes and suite of advanced sensors. In that moment of truth before the collision, should the vehicle target a small car, rather than a big one, to err towards protecting its master? Or should it do the reverse, aiming for the SUV, even if it means reducing the robo-car owner’s chances of survival? And what if it’s a choice between driving into a school bus, or plowing into a tree? Does the robot choose a massacre, or a betrayal? (www.popularscience.com)

Patrick Lin, writing in Wired (5.6.14) has written a seminal piece on ethics and robotics.  He suggests that the crash-optimization ethical choice – i.e. programming the smart car to first save itself and its passengers – may be unethical.  However in scenarios where only others are at risk, what actions should the car take?

The problem is starkly highlighted by the [following] scenario, also discussed by Noah Goodall, a research scientist at the Virginia Center for Transportation Innovation and Research…Imagine that an autonomous car is facing an imminent crash. It could select one of two targets to swerve into: either a motorcyclist who is wearing a helmet, or a motorcyclist who is not. What’s the right way to program the car?

In the name of crash-optimization, you should program the car to crash into whatever can best survive the collision… Here, it means striking the motorcyclist who’s wearing a helmet. A good algorithm would account for the much-higher statistical odds that the biker without a helmet would die, and surely killing someone is one of the worst things auto manufacturers desperately want to avoid.

By deliberately crashing into that motorcyclist, we are in effect penalizing him or her for being responsible, for wearing a helmet. Meanwhile, we are giving the other motorcyclist a free pass, even though that person is much less responsible for not wearing a helmet, which is illegal in most U.S. states.

Perhaps new owners of smart cars will be able to select from a number of ethical options.  Some may choose Option A according to which the life of the owner comes last, and the car should take evasive actions which will first avoid death or injury of others. Swerving over a cliff, for example, to avoid both cyclists.  Option C (1) – the so-called ‘Jain Option’ - would avoid killing any animal even if it means serious damage to the vehicle.


Ethics aside, another principal issue concerning automated driving is legal liability.  If the smart car in which you are driving kills a cyclist, as in the example above, who is responsible? The owner of the car who was playing cards in the back seat? The car manufacturer? Google?



Eventually the car itself will assume legal responsibility, but not until artificial intelligence reaches a point where it is more intelligent than the ‘real’ version. If the smart car program of the future always makes rational, unerringly correct driving decisions; if it is self-correcting to that all minor glitches and tics are automatically fixed; and if it can replicate itself in every-improving versions, then it will be considered truly autonomous, ethical, and legally liable.

For time being consumers would have to choose the ethical and legal program options best suited for them. Option C would program the car to take only those actions which limit legal liability.  Option C (1) would balance ethics and liability.  Option C (2) would always place ethics first.

Automated cars will involve far more than mechanistic solutions – avoiding telephone poles, measuring 12’ from a fire hydrant, coming to a complete stop at stop signs. They must act human and therefore make appropriate ethical and legal choices.

This is why the smart, automated car is such a revolutionary invention. Robots with a rivet gun on an assembly line pose no challenges.  Nor do smart vacuum cleaners, self-replenishing refrigerators, or fully automated greenhouses. 

A smart car, however must have programs so intelligent that they can mimic the most sophisticated human decisions.  Once Google has crossed this threshold, truly intelligent robots which can exist side-by-side with us, be our co-workers, mates, and partners will not be far behind.