Engineering Ethical Study - University of Pittsburgh

advertisement
Vidic, 2:00
R09
A SAFETY ETHICAL DILEMA
Mason Unger (mhu3@pitt.edu)
data and possibly lose my job. I decided to consult codes of
ethics, similar case studies, and consumer opinions to see if
they could give me any advice in this situation.
MY CASE
My name is Mason Unger, and I am a software engineer
at Google. I have worked on Google’s self-driving car project
for the past four years. I have been tasked with researching
the safety of our vehicles. I often ride in our test vehicles and
monitor how our cars make decisions; then based on my
observation, modify programs to optimize safety. Over the
past two years, our cars have driven over two million miles.
Over that span, I have recorded 20 accidents in total. Most of
these accidents were deemed to be the fault of our cars. Most
were due to people not paying attention (i.e. using handheld
devices) or people being so amazed by our car that they lost
focus and crashed into our vehicle. Although, one accident
left our test driver in the hospital.
Our car was driving down the road when a child ran out in
front of it. To avoid the child the car engaged its breaks and
the car following (too close) behind it swerved into oncoming
traffic. This caused an oncoming car to swerve into the
google car. Another four situations like this one occurred
leaving our test drivers injured but not hospitalized. In
summary, our cars have experienced twenty crashes. Five of
these crashes posed a serious threat to our drivers, but the
other fifteen would be classified as minor crashes with little
risk of injury.
After reporting this at our latest safety board meeting, I
was called into Chris Urmson’s, head of the self-driving
program, office. Here we talked about the upcoming Google
unveil (where we would be presenting our self-driving cars
and all information associated with them. Urmson saw safety
as the number one concern of the public. He feared that if the
public deemed our car unsafe, we would lose interest, and the
program could be terminated. He then told me not to report
the five major accidents.
The next day, I was having second thoughts about
withholding information from the public, even if it benefited
our project. I talked to Urmson again. I stated that even if we
report all twenty accidents, that is still only one accident for
every 100, 000 miles, which is exactly the average for
American drivers. In a sense, our cars are just as safe as
conventional drivers. Urmson stated that the five major
accidents would scare the public. He thought that even one
life threatening accident would throw the publics opinion. I
stated that we only needed more time to optimize the system.
After all, the major accidents occurred within the first year of
test driving. Urmson stated that he had promised that our selfdriving cars would be consumer ready within the next year,
and a setback like this would make it impossible for our cars
to be consumer ready. In order to make the deadline, I was to
falsify our safety data. I had a choice to make. Either falsify
the data and endanger human lives, or present the true
CODES OF ETHICS
Codes of ethics are sets of rules to govern all disciplines
of engineering. These codes are helpful to engineers when
ethical decisions must be made. The National Society of
Professional Engineers (NSPE) has a code of ethics that
governs all engineering disciplines.
The first fundamental canon states that engineers must
always hold the safety, health, and welfare of the public above
all else [4]. In my situation, If I were to lie to the public and
our cars would to go onto the road, I would be putting the
public’s safety at risk. A by-law of the first cannon states that
if an engineer’s judgement on the public’s safety is overruled,
he or she should notify proper authorities [4]. My judgment
as an engineer was overruled when Urmson stated that the
cars would be safe enough. This by-law encourages me that
I am right, and I must find a higher authority. Because of this,
I now have a third option; I could go to Google’s CEO, Sergey
Brin and tell him the situation. A second by-law states that
engineers can not reveal information without the consent of
their employer, unless that information is required by the code
[4]. This statement puts me in a tough spot. The only way I
could reveal the information is if I was given consent to do so,
or if withholding the information is in violation of the code.
The third canon in the NSPE code of ethics states that all
public statements must be objective and truthful [4]. This
proves that withholding information is against the code. If I
were to give the public false information, I would be violating
the third cannon. This gives me the right to present the public
with my true information because doing otherwise would be
a violation. The NSPE code of ethics shows that keeping the
information from the public is immoral, but it does not help
me analyze the consequences of this. If I present my true
findings, the public will likely abandon the Idea of selfdriving cars. This would shut our project down and my coworkers would all lose their jobs. This is just a theory of
mine, so I will continue to analyze sources that will help me
make the best decision.
The second code of ethics that I chose to consult was the
Software Engineering (SE) code of ethics. This code directly
governs my branch of engineering. One key point in this code
states that software engineers must take full responsibility of
their work [6]. This means that I am directly responsible for
the safety of the google cars. This also implies that I am
responsible for the safety information, and on that note,
responsible for telling the truth in matters regarding the safety
of our cars.
1
University of Pittsburgh, Swanson School of Engineering
2015-11-03
Mason Unger
The code also reads, “Software engineers shall approve
software only if they have a well-founded belief that it is safe,
meets specifications, passes appropriate tests, and does not
diminish quality of life, diminish privacy or harm the
environment. The ultimate effect of the work should be to the
public good [6]”. Obviously our cars need more testing. As
a software developer and safety manager I cannot let an
unsafe product put innocent people’s lives in danger. If I were
to present false information, our cars would become
commercially available.
The code states that I must disclose to appropriate persons
any actual or potential danger to the public, and also must
avoid deception when addressing the public [6]. I have
already addressed my concern to the leader of our self-driving
project, Chris Urmson, but he is the person who told me to
falsify my records. This would mean that the next person I
would have to speak to is Google’s CEO, Sergey Brin.
According to the SE code of ethics, it would be immoral for
me to present falsified information to the public.
These codes have helped me determine the morality of one
of my choices, but I must consult other sources before I make
a decision. It is clear that engineering codes of ethics prove
that it is unethical to falsify information that would harm the
public. However, it may be immoral, but it may not be wrong.
I also must weigh the repercussions on me, my co-workers,
and the company before I make a decision.
production, the cars may still have some bugs that cause
serious accidents. But like in the case study, if we show that
our cars are unsafe, we may lose interest, funding, and our
jobs. It would seem that if we could correct the problem
before the car was made commercially available, this could
be a positive situation just like the case study.
PUBLIC OPINION
In order to see this dilemma from a different perspective,
I decided to talk to two people. The first person, Justin
Shaffer, is an avid car lover [2]. He works as a mechanic, and
has been working with cars for the past thirty years [2]. I
wanted to ask Justin what he thought about self-driving cars
and their safety.
He stated that he would trust a computer to drive and avoid
accidents more than he would trust himself [2]. This was very
shocking to me. I asked him what if these cars got into just as
many accidents as the human driver. He stated, “The
positives from not having to drive outweigh the negatives
[2]”. As a car enthusiast, I expected Justin to be completely
against self-driving cars. His only concern is that he would
not buy a self-driving car unless it had a steering wheel
because being able to drive when he wants is very important
[2]. So if a conservative car lover would not be phased by
self-driving cars, what would the general public think?
To answer this question I decided to talk to my friend,
Ashley Shaffer. Ashley has earned her master’s in social
work [1]. She works as a social work researcher at
Shippensburg University [1]. Her schooling and research has
led her to become very good at understanding the public and
their actions.
I asked her what she thought the public would think if I
presented all of the safety data. She stated, “The public would
react poorly. Who would want to get in an accident that they
had no control in? Even though it’s not true, most people
think that if they are in control, they can prevent harm [1].” I
then proceeded to ask her what she thought I should do. She
stated that I should go to a higher authority, and if Google’s
CEO still wants me to present the false data, I should listen to
him because then I are no longer as liable for the information
[1]. She also recommended that I look at other ethical sources
that were not engineering related, but instead analyzed ethical
dilemmas holistically [1].
SIMILAR STUDY
A case study that closely mirrors my dilemma was
recently written by Michael McFarland. In his case study,
Wayne Davidson, a software engineer for an aerospace
engineering firm, is in charge of determining the safety of a
flight tracking program [3]. He deems the program to be
unsafe for commercial use because it often loses track of
planes (this would result in crashes and put the public at risk),
but his boss urges him that the company purchasing the
software (The Federal Aviation Agency) will run many test
before it is actually in use [3].
This gives Wayne’s company time to fix their program
and update the software for the FAA before it begins to use it
large scale. His boss claims that because the FAA will only
be running tests and small scale simulations for the next
month, the program is safe [3]. Also, if Wayne’s company
does not present the program, they will lose their contract with
the FAA, and thousands of jobs will be lost [3]. Ultimately
Wayne chose to present the software as is and then later fix it
[3]. This caused no problems and the software was fixed
before it was used large scale [3].
In my situation, I know that the software used to control
our self-driving cars is unsafe. My boss has asked me to give
false information that shows our cars to be safe. Although in
my case, I may not have enough time to completely make the
cars safe before Google’s promised launch date. We could
continue to update the vehicles, but for the first year of
ANOTHER ETHICAL SYSTEM
To analyze this problem from a wider scope, I decided to
consult a code of ethics that is not associated with
engineering. I chose to look at Utilitarianism, and in specific
Jeremy Bentham. Bentham was a British philosopher in the
late 1700’s [7]. I chose to apply Bentham’s ethical system
because it seems to fit my dilemma.
Bentham’s ethics are all based on his greatest happiness
principle [7]. This states that the best course of action causes
2
Mason Unger
the most happiness and the least pain [7]. The only way to
determine the correct action is to weigh the happiness of
everyone involved [5]. This is called felicific calculus or
hedonistic calculus [5]. In order to use this system, you must
understand the seven areas that get analyzed for each person
or persons. The seven areas are intensity (how intense is the
pleasure/pain?), duration (how long does it last?), certainty
(how certain are you that this will cause the pleasure/pain?),
propinquity (how soon will the pain/pleasure occur?),
fecundity (will the pleasure/pain lead to other
pleasures/pains?), purity (is the pleasure free of pain?), and
extent (how many people are effected?) [7]. All of these
values can be assessed on a scale from -10 (pain) to 10
(pleasure) [5]. The action that has the highest overall score is
the action that should be pursued according to Bentham [5].
I decided to analyze the three choices that I had: to tell the
public the truth, to lie to the public, or to go to Google’s CEO,
Sergey Brin. After using Bentham’s felicific calculus, it is
obvious that talking to Sergey Brin should be my course of
action. In the other first scenario, more people were harmed
by the truth, even though in telling the public the truth, I was
following the codes of ethics. Also in this case the public does
not gain any happiness. In the second scenario, the lies caused
harm because of the number of consumers effected, and the
seriousness of the harm. Going to Sergey Brin does not cause
pain or pleasure, so in this case, it is the right course of action
according to Bentham.
RECOMMENDATION
Engineering is full of ethical decisions. When faced with
a dilemma engineers must consult as many sources as
possible. Codes of ethics can be a great start, but just because
an action is unethical according to the code, does not make it
wrong. You must factor in the people effected and the effects
of your actions. This could be the loss of funding, jobs, or
even lives. Engineers must also consult a variety of sources.
If you only consult engineering based sources, you may come
to a different conclusion. I recommend that you look at public
opinions. The reason we are doing engineering projects is to
help the public, so it would make sense to consult your
consumers. Finally, an engineer faced with an ethical
dilemma should consult ethical systems that are not related to
engineering. This can give a broader view of the whole
situation. By consulting a variety of sources one can make
the correct decision, and as engineers, it is our duty to make
the correct decision.
REFERENECES
[1] A. Shaffer. (2015, October 30). Interview.
[2] J. Shaffer. (2015, October 28). Interview.
[3] Michael McFarland. (2012, June). “Occidental
Engineering Case Study: Part 1.” Markkula Center for
Applied
Ethics.
(Online
article).
http://www.onlineethics.org/Resources/Cases/OccidentalEng
.aspx
[4] “National Society of Proffesional Engineers Code of
Ethics.” (N.D.). National Society of Proffesional Engineers.
(Online article). http://www.nspe.org/resources/ethics/codeethics
[5] “Philosophy 302: Ethics, The Hedonistic Calculus.”
(N.D.).
Lander
Philosophy.
(Online
Article).
http://philosophy.lander.edu/ethics/calculus.html
[6] “Softw are Engineering Code of Ethics” (N.D.). Coputer
Engineering
Society.
http://www.computer.org/web/education/code-of-ethics
[7] Wesley C. Mitchell. (1918, June). “Bentham’s Felicific
Calculus.”
Academy
of
Political
Science. http://www.jstor.org/stable/2141580
CONCLUSION
After reviewing engineering codes of ethics, case studies,
consumer opinions, and an ethical system outside of
engineering, it is clear that I should seek out Sergey Brin. The
codes of ethics support this decision because they both
directly state that I must disclose potential danger to the public
to the appropriate people (in this case, Brin)[4][6]. In a
similar case study, a software engineer was faced with a
similar dilemma. His program was deemed unsafe, but he had
enough time to fix it, so he decided to sell the unsafe software
in hopes of fixing it later [3]. This saved many jobs [3]. In
my case, I may not have enough time to fix the safety issues
associated with our car, so it is logical that I should go to
someone in charge. One person that I Interviewed, Justin
Shaffer, stated that he would still feel safe in a google car [2].
The second person, Ashley Shaffer, stated that the public
would react poorly to the safety report, and also stated that I
should talk to Google’s CEO [1]. Finally, Jeremy Bentham’s
felicific calculus proves from a different perspective that
speaking to Sergey Brin about the safety data is the correct
choice [5][7]. After meeting with Brin, he decided that we
should postpone our release date and the date of our
presentation until we can make our cars more consumer
friendly.
OTHER SOURCES
“Air bags, Safety, and Social Experiments.” (N.D.) Online
Ethics
Center.
(Online
Article).
http://www.onlineethics.org/Resources/Cases/AirBag.aspx
“Obligation to Client of Employer?” Online Ethics Center.
(Online
Article).
http://www.onlineethics.org/Resources/Cases/Obligation.asp
x
3
Mason Unger
ACKNOWLEGEMENTS
I would like to thank Anjali Sachdeva, Ashley Shaffer,
Ethan Henderson, and Nowa Broner for encouragement and
advice.
4
Download