Sei sulla pagina 1di 5

Ethics of technology and science

Charalampos Orfanidis
December 7, 2016

Describe shortly and compare between the princi-


ples and rules of VR, ICSU and CUDOS. Explain
what are the main common grounds for all these
codices, and identify possible conflicts between
these ethical guidelines
VR, ICSU and CUDOS are three different frameworks which define a series of
ethical standards regarding science.
VR propose a set of rules proposed from the societys point of view in order
to ensure the high expectations of it. As a scientist is conducting research he
has to conduct it responsibly towards the people and animals which are involved
and even more responsibly to the ones (people, animals environment) that are
affected indirectly. High quality research is characterised as a result that comes
from scientists that are not forced from external factors or when it is not coming
from personal interests of the involved parties. One of the characteristics of a
successful scientist is the trust from the society. Some of the principles enforced
from VR are the following: Being honest about your research, do not external
factors leverage your research purposes, publish your methods and results than
can help others (scientific community, society or industry), do not steal research
findings from other scientists, organize and document your research in order to
be accessible to others, make sure that your research do not harm other people,
animals or environment and be fair when reviewing others research.
ICSU is introducing a set of ethical standards from the society point of view
as well but with an important constraint. They mention that asking scientists to
conduct their science responsibly considering the society it may give them power
in a group of people that do not know how to cope with it. It is proposed when
there defined ethical standards by scientists they must do it in an open manner
and include people outside from their research. Another issue they try to tackle
is that ethics are defined different from country to country and it is important
to adopt an international code of ethics in science. To this end they focus on
three subjects namely: freedom and responsibility in science, animal welfare,
science in the internet era. The first one, freedom and responsibility in science

1
analyses how freedom in science may be limited by various factors such as reli-
gion, politics, external interests. Regarding the responsibilities are divided into
internal for their discipline, profession or colleagues and external for the society.
The subject including the animal welfare outlines the necessity of considering
the standards of animal experimentation and how non-human objects can also
be moral subjects. The section about science in the internet era mentions some
security measures that countries countries with vulnerable infrastructure should
take. Moreover it focus on the public sharing of the academic literature how
different terms apply on this issue under different societies.
CUDOS framework is based on three academic values Communism, Univer-
salism and Disinterestedness. Communism is presented as a principle related to
public sharing of research results and teaching materials instead of concealing
research material with copyrights and restricting access. Universalism is intro-
duced as a value which will be able to overcome national, political and religious
limitations and have objective results which the aforementioned limitation will
not be able to affect them and vary when these are different. Disinterested-
ness according to CUDOS is a principle which drives the scientist to conduct
his/her research with only scope to find the truth. Every other motivation which
can mislead the scientist from this path such as emotional, economic, career is
wrong.
If we try to compare the previous three frameworks we can find many sim-
ilarities and differences as well. VR and ICSU have in common the principle
which protects and benefits the society from research and scientists should be
responsible of it. ICSU and CUDOS have both the principal of public sharing
the research results/literature and try to outline how this can be beneficial. One
difference is that according to CUDOS a scientist has to research in order to
find the truth and not be affected from any external emotion but if a researcher
during his/her research finds out that his/her finding can be used to create
weapons and wants to change his direction or destroy the results this is what
the first two frameworks propose but according to CUDOS the scientist has to
disclose the results.

2
Describe an ethical problem from your area of
research and identify which principles from the
codices of VR, ICSU and CUDOS are relevant
for it. Try to solve the problem using these prin-
ciples, and try to find conflicts between the prin-
ciples which are applicable in this situation. Pro-
pose a solution and discuss what supports and
what opposes your solution with regard to the
principles and rules you refer to
An ethical problem in engineering is the gender equality. Most of the fields of
engineering are dominated by male scientists such as computer science, mechan-
ical engineering. This has as an effect numerous problems for female scientists,
for instance being difficult for a woman to get a faculty position regardless her
technical skills or scientific papers which have first author a female name getting
worse reviews than others that have first author male names.
According to VR the last proposed principle is the following: be fair in
your judgement of others research.. In the problem mentioned above when the
male reviewers review with bias the paper the violate this principle. Also in
ICSU it is mentioned in the first subject of rules, freedom and responsibility in
science, that there are some internal responsibilities regarding their discipline
their profession and their colleagues. If there is a discrimination between male
and female scientists this principle is clearly violated. If the researches try to
apply these principles the gender inequality will not be a problem any more.
This inequality problems were even more intense in the past but they occur
even today and the scientific community struggles to reduce it with several
methods. Some ways to mitigate these phenomena are conferences devoted to
women engineers such as womEN encourage, which is organized by Association
Computing Machinery. write about this conference. Another way to improve the
situation is to enforce international regulations of having a certain percentage of
gender equality in big technology companies and academia. The later strategy
is used as a pilot program in some companies in EU. Another tactic to have
a transparency in the review process of scientific papers is the double blind
process which means that both the reviewer and author identities are concealed
from the reviewers, and vice versa. This is the best solution for having a fair
review and tackles one of the mentioned problems but is not followed from all
the journals. In order to have a universal transparency we have to force all the
journals follow this strategy.
Although the above solutions may lead to some conflicts because in the
attempt to have a gender equality in a technology company the manager may
a hire a female engineer instead of a male engineer with better technical skills
just to adhere to the gender equality regulations. This is also a wrong strategy

3
and unethical as well.
My personal point of view for tackling this problem is that the study pro-
grams (especially the ones that are dominated by male students) should be
taught these issues and be aware of this problem from the very first years of
their studies. Since it is easier to tackle a problem like this from the very
beginning instead of trying to persuade a senior scientist which might be bi-
ased already, that gender equality is important. Also it is important to enforce
seminars explaining this issue, and force to follow as many members of the com-
munity as possible because there many that act unconsciously and do not how
they harm the career of a female scientist.

Essay
Recently a article from MIT technology review gain publicity entitled How to
Fix Silicon Valleys Sexist Algorithms [1], the problem here is that we teach
inadvertently an Artificial Intelligence (AI) system to be sexist. More specifi-
cally a new research shows that gender bias is conveyed through data sets used
to teach language skills in AI systems. As AI systems are integrated more
and more into technology, their gender biased perception may have unexpected
negative effects such as misguided results in job searches. The problem arises
from the training data that are fed to these systems which are mostly written
or spoken language. In this way the AI systems are creating some relations
between words and phrases. Scientists from Boston University and Microsoft
Research new England found some interesting results about the word man
and woman. The first word seem to have a closer relation to the word pro-
grammer and the second a closer relation to the word homemaker.
Another scientist from Stanford University made some experiments and con-
firmed that if you use this system for recruiting purposes for programmers, there
more possibilities to be proposed male candidates. This research group tried to
eliminate this bias by modifying the algorithm of the AI system but there were
some scientists claiming if we modify the algorithm then the AI system does
capture the real word image but something artificial.
Barbara Grosz, a professor from Harvard University comments When you
are in a society that is evolving in certain ways, then you are actually trying to
change the future to be not like the past and then she mentions that there is an
ethical question if we are changing the evolution in a path that we want.
Lets examine this ethical dilemma by using the three different frameworks
we used in the above questions. The ethical dilemma here is if we should modify
the AI system in order to remove the gender bias from its given results. Ac-
cording to VR you have to tell the truth about your research and you should
strive to conduct your research without harming people. These are the most
applicable rules. For the first if someone presents every detail of the system it is
ok, but for the second if we consider that the gender bias can lead to any kind
of harming people then the gender bias should be removed from the system.
However there is a comment in the article from the professor which supports

4
we should not remove the bias, that this gender bias it really depends how it
affects the results regarding the application it is being used. This makes sense,
because if you want to use the AI system for depicting the real society point of
view you have to include the gender bias in system. For other cases such as the
job searching example it would make sense to remove it.
If we try to follow the ICSU framework, there is a set of principles include
in the freedom and responsibility in science section, there are mentioned some
internal and external responsibilities, the former regarding their profession, dis-
cipline and their colleagues and the latter regarding to the affection of the
research to the society. This ethical dilemma falls within both internal and ex-
ternal because it can affect the female colleagues of the discipline and it can also
affect the society by forwarding this sexual inequality point view in many levels.
Again it depends how the AI system is used and under different circumstances
a responsible scientist can decide if this feature has to be removed or not.
Using as a guide the CUDOS framework we may have some different deci-
sions. For instance if we apply the disinterestedness principle which is defined
in order to motivate researchers to conduct their research with only motivation
to discover the truth, then if we remove the gender bias from the AI system is
violating this principle because it does not represent the truth anymore. If we
elaborate on this we can think that is another principle on CUDOS framework
called Universalism which is used to avoid different kind of limitations (political,
religious, national) to have objective results. This principle enforce the opinion
to not remove the gender bias because the biased results will be generated from
the constructed system which depicts the objective perception of the society. It
is the society itself which includes and reproduces the gender bias stereotype.
Summarizing we can say that deciding regarding the application if should
include this feature or not seems to be the more ethical and logical decision.
Sometimes it can be useful but some others can lead to misleading results. It
is important to focus in the ethical question of if we changing the direction
of the evolution in the way we want or not and if we do so, what are the
consequences. Another interesting question on this is how much we know that
the chosen direction is the correct one. These all are questions that have more
than one answers and should be examined thoroughly because AI is a field that
can affect society in manners that we do not know today. Recently Elon Musk,
famous entrepreneur and engineer, expressed his concerns about how AI and
how dangerous can it be for humanity. Furthermore he founded a non profit
company in order to investigate how dangerous can it be in the futures and
keep humanity safe from it. This now may sound like a scenario from a science
fiction movie, but these are real facts which remind us that he have to be aware
and handle these issues properly.

References
[1] How to Fix Silicon Valleys Sexist Algorithms,
https://www.technologyreview.com/s/602950/how-to-fix-silicon-valleys-sexist-algorithms/

Potrebbero piacerti anche