The Faults In Our (Self-Driving) Cars

Logan Soss

Mrs. Schebesta

English 9A Period 4

Friday, May 15, 2015

​ ​
Depictions of the future in movies and literature would not be complete without some

form of robots in control of our transportation. Although these same stories create warnings

against having androids rule critical aspects of society, they fail to recognize that driving is also

one of the most dangerous and complex everyday tasks. Driving requires a vehicle’s operator to

make sudden ethical choices in emergency situations. Some of these choices are lose-lose, like

in life-or-death scenarios, where the best outcome is controversial and depends on the viewer's

moral standpoint. Because of this, the way to react cannot be calculated and pre-programmed

into the robot as predetermined choices. As sensors improve, robotic technology should

continue to provide assistance to a human driver. But no matter how advanced, robots can never

replace the decision making needed to drive the vehicle.

The current state of technology is far behind what is needed to safely and autonomously

drive a vehicle. Google’s self-driving prototype still relies heavily upon maps, which can create

some problems (Griffiths 1). Since not all of the roads in the world are mapped, the vehicle may

not be able to go on private or foreign roads. Running a car on a crowd-sourced network can

also lead to false information of traffic and closed roads being received by the car. Today's

detection technology is not perfect, and sensors cannot detect potholes or other road damage.

Griffiths explains that the cameras on Google’s self-driving car may become impaired if struck

directly by a beam of sunlight, and the vehicle “can’t cope in snowy conditions and cannot

negotiate heavy rain” (2). Even if the sensor technology were to improve to a state that they

could reliably feed information to the vehicle, the network system itself is very susceptible to

hacking.
 Soss 2

Self-driving cars rely on a vehicle-to-vehicle network, communicating to each other by

sending information as a way to prevent accidents from happening. Because of the wide variety

of vehicles and the need for all cars to share information, the network needs to be open and

easily accessible. But this openness also leads to security issues, where a hacker could easily

take control of the car. Hacking of cars is already possible now and has happened on

non-autonomous cars, “In 2011, a different team of researchers managed to penetrate similar

systems through bluetooth, mobile data, and even a malicious audio file burned onto a CD

played in the car’s media player” (Hern 1). As developments in technology continue to grow,

there will always be third-party sources that take advantage of the system for their own causes.

The supporters of self-driving cars trust the technology today. It is a fact that “research

shows that 90% of crashes are caused by human error” (“The Road To…” 2). Cars today are

already equipped with driving-assist features like automatic braking, cruise control, and parking

assist. Supporters argue that these advancements are already present, and improving upon them

can help to eliminate the 90% of accidents caused by human error. While it is true that these

advances can continue to improve and help with the technical side of driving, what supporters

fail to recognize is that there are ethical issues and concerns that need to also be addressed.

Jason Millar created a question known as the Tunnel Problem, an ethical experiment still

studied today. The reader is presented with the scenario that they are in a self-driving car, moving

down a single lane road towards a tunnel. Suddenly, a child dashes into the road, trips and falls,

blocking the car’s path. There are only two options: swerve into the side of the tunnel, missing

the child, and in turn killing themselves, or staying on the road, saving themselves but killing the

child. A website, Robohub, took this question and created a poll asking similar questions, and
 Soss 3

the results are shocking. 36% of the voters chose to swerve and kill themselves, while 64%

chose to hit the child. When the viewers were asked how hard it was to answer the question,

nearly half of them (48%) said that it was an easy question to answer, with 28% saying it was

moderately difficult, and only 24% saying it was difficult. At the end, Robohub asked one final

multiple choice question: should the passenger, lawmakers, or vehicle manufacturers and

designers determine how the car responds to the environment?. (“If Death By…”) This poll

raises an ethical dilemma, and Patrick Lin basically sums up the whole problem in one sentence,

“damned if you do, damned if you don’t” (1). One given solution to the problem is to apply a

ethical setting which can be set for each passenger.. This will let the passenger of the car set the

value of their own life relative to the ones of others around them (Lin 1). Unfortunately, even if

an ethics setting was provided in autonomous vehicles, there would be no way for a computer to

analyze all the factors needed to make the moral decision, such as whether a 70 year-old man

falls in front of the car versus a 6 year-old girl. And if there are multiple passengers in the car,

each passenger may have a different desired setting, making this solution not very practical.

The idea of a self-driving car has brought up some other issues as well. Since

autonomous cars will have to deal with all the choices and knowledge a driver has to have, the

vehicle will need to deal with many variables. As it would take a very long time to test every

single one of the million factor combinations, there are bound to be bugs and glitches in the

system which can create unexpected results.

Besides the ethical decisions faced by driving a vehicle, there are also the social concerns

brought about by displacing humans by autonomous drivers. Uber is an app where a user types

in their location and destination and are provided with drivers and estimated prices. A couple of
 Soss 4

days after Google released their self-driving prototype, Travis Kalanick, Uber’s co founder, said

that hopefully one day in the future, their company can use driverless vehicles instead of

humans. With this bold move, the prices for rides will decline because the ‘middle man’ (the

driver) will be eliminated. But of course, with Kalanick’s statement, all of the drivers would be

laid off, which sparked some protests by Uber’s drivers. (Newton) With more than 4 million

professional drivers today (in 2011, there were about 233,000 taxi drivers and chauffeurs, about

3.5 million truck drivers) all of them could be left unemployed by autonomous driving vehicles

(“Occupational Outlook Handbook…” 1) (“Truck Drivers in...” 1).

Today’s technology is far from capable to safely drive a vehicle. Even with advances in

sensors and programming, the ethical decisions required do not have an automated solution.

Driving requires the driver to make quick, moral choices, and the only way for robots to mimic a

human is to be fully self-aware, which leads to a new set of ethical questions. While advances in

technology can continue to make a driver’s job easier, it cannot ever replace the role of a human

as the driver.
 Soss 5

Works Cited

Griffiths, Sarah. "Is Google's Driverless Car a Fair Weather Friend? Vehicle WON'T Work in

​ Heavy Rain or Snow, Insider Reveals." Mail Online. Associated Newspapers, 1 Sept.

​ 2014. Web. 22 Apr. 2015.

Hern, Alex. "Self-driving Cars Irresistible to Hackers, Warns Security Executive." The

​ Guardian. 28 Jan. 2014. Web. 22 Apr. 2015.

"If Death by Autonomous Car Is Unavoidable, Who Should Die? Reader Poll Results." Robohub.

​ 23 June 2014. Web. 29 Apr. 2015.

Lin, Patrick. "Headline: Here’s a Terrible Idea: Robot Cars With Adjustable Ethics Settings."

​ [Link]. Conde Nast Digital, 18 Aug. 2014. Web. 29 Apr. 2015.

Newton, Casey. "Uber Will Eventually Replace All Its Drivers with Self-driving Cars." The

​ Verge. 28 May 2014. Web. 26 Apr. 2015.

"Occupational Outlook Handbook: Taxi Drivers and Chauffeurs." U.S. Bureau of Labor

​ Statistics. 8 Jan. 2014. Web. 6 May 2015.

"The Road to Self-Driving Cars." Consumer Reports. 1 Feb. 2014. Web.