IEEE Technology and Society Magazine - March 2021 - 88

AI for a Fair, Just, and Equitable World

Social impact advisor and measurement system
(SIAMES) seeks to support the measurement of
social, economic, and environmental impacts, especially in the field of social economy. The measurement of social impact, though a complex job, can be
carried out on any project in the social field, providing objective and truthful information on the results
of the different actions, thus improving the social
return on investments (SROIs) made.
With SIAMES, we pursue applying ontology-based
semantic text mining methods on a corpus of impact
measurement reports to subsequently use the results
in a recommender system applying case-based
reasoning (CBR).
To do this, we will define an initial ontology,
which we will populate through text mining. Subsequently, we will create a CBR system to retrieve indicators for specific social entrepreneurship projects.
The ontology we have developed is of special
importance since, as far as we know, it is the first formal knowledge model in the field of social impact
measurement. This ontology not only will serve
for the present research, but also could constitute
a starting point for different subsequent work in
this area.
It should be noted that the measurement of social
impact is not limited to the scope of social enterprises. Several agents may need to measure their
social impact, perhaps the most evident being is the
public sector, which needs to strive for efficiency. To
ensure the proper functioning of the policies carried
out, the impact of projects implemented directly by
the public administration could be measured without too much difficulty by adapting the approaches
proposed for social enterprises.
Moreover, there is a trend for traditional companies to integrate corporate social responsibility
(CSR) policies into their business management
processes. Since CSR is controversial [8], impact
measurement [9] of these policies is desirable. The
artificial intelligence (AI) sector stands out, where
because of the current debate on the ethical risks
involved, integrating impact measurement modules
in the applications themselves [10] is under study.
Finally, the need for suitable indicators for the SDGs
has also been highlighted [11].
Therefore, the development of an ontology on the
measurement of social impact is particularly relevant
given that we did not find any reference in this field

88

in our bibliographic search, although, however, we
have found some relevant related work [12], [13].
In this short article, we will only show our research
preliminary results and explain the difficulties we
faced in this task. These results consist mainly a first
approach to an ontological similarity function to be
used by the recommendation mechanism.

Ethical implications
As it has been noted in the introduction, the execution of social projects, especially those related in
one way or the other to the business environment,
can generate controversy. For instance, the surprising positioning of some companies, assimilated by
the public as unethical, in certain social causes, has
coined terms such as pinkwashing4 or greenwashing
[14], referring to this alleged image washing.
On the other hand, social entrepreneurship itself
also has its detractors. The tendency to take for
granted the interest for the common good of these
initiatives is common; for this reason, it is convenient not to lose the critical spirit to avoid losing,
consequently, the social focus. Hyunbae's article
Politics, Values, and Social Entrepreneurship: A Critical Appraisal [15] exposes the most controversial
aspects in relation to social entrepreneurship; knowing and keeping them in mind is an essential part
when working in this area.
SIAMES, being a tool to support social projects,
would also be, immediately and for obvious reasons,
socially positive. However, what happens in these
cases where the social aspect of the initiatives is not
so clear?
The subjective vision of human beings allows us
on many occasions to discern whether, in our perception, the objective behind an initiative that calls
itself social really has a positive impact on general
well-being. AI tools aimed at measuring positive
social impact, unless they have specific implementations for it, will not be able to evaluate the real
purpose in such an abstract way, and may even
favor initiatives that meet certain criteria objectives
despite being far from being truly social initiatives.
In addition to these types of conflicts, related to
actions by individuals or companies carried out in a
conscious and ethically reprehensible way, we find
that AI can also make mistakes that unpredictably
cause unfair situations.
4
https://www.nytimes.com/2011/11/23/opinion/pinkwashing-and-israels-use-ofgays-as-a-messaging-tool.htm

IEEE Technology and Society Magazine
https://www.nytimes.com/2011/11/23/opinion/pinkwashing-and-israels-use-of-gays-as-a-messaging-tool.html https://www.nytimes.com/2011/11/23/opinion/pinkwashing-and-israels-use-of-gays-as-a-messaging-tool.html
IEEE Technology and Society Magazine - March 2021

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2021
