IEEE Technology and Society Magazine - March 2021 - 59

large volumes of data that are used to dynamically
balance the load of the system according to the needs
in a more efficient way. In the case of smart grids, not
only do they transport energy but also information, as
in the case of volatile energy production plants such
as solar and wind power, it is increasingly vital that the
production and distribution of energy smartly react to
consumption. With the help of AI, the information
of the different stakeholders involved in the network
(producers, storekeepers, and consumers) can be
evaluated, analyzed, and monitored. In the case of
volatile energy production, AI is also being used to
anticipate production with predictive techniques that
allow forecasting the future produced energy to plan
the load on the network [39], [40]. One of the biggest
problems in resource distribution networks infrastructure is maintenance. It is very complicated to manually monitor hundreds of thousands of kilometers of
pipes or cables, to detect problems, leaks, or accidents that can have serious consequences such as
fires or floods. The digitization of networks using IoT
allows the collection of large volumes of information
about the state of the networks and the service provision that can only be handled with high-performance
AI-based approaches, providing valuable information
almost in real time, allowing a much safer and more
efficient maintenance [41], [42].
AI has also proved to be a valuable asset in the
management of water quality for human consumption. The digitization of the river basins provides realtime information about the chemical composition of
the water and its volume, which can vary depending
on the climate and the industrial waste discharges,
known or unknown. This information can be conveniently processed using ML techniques, allowing
safer and more efficient management of the wastewater treatment plants [43].

Ethics and AI
Ethics is a broad concept of philosophy that seek
to answer questions regarding human morality that
aims to determine what are considered good and
bad actions. One of the main challenges of today's
globalized world is to define a group of ethical
values or norms that reach a general acceptance,
due to social, economic, and cultural differences
among different regions. Today, our society is ruled
by a set of laws defined either at a primary level,
such as the Universal Declaration of Human Rights,
and a secondary level with national or international

March 2021

laws (e.g., General Data Protection Regulation of
the European Union). However, regulations are, in
general, a slow process that cannot follow the pace
of technology progress. Therefore, AI is not legally
bonded to any rule, although some aspects can be
found in domain-specific laws, as it is the case of the
European Medical Device regulation for the health
care sector [44]. Hence, AI design, development,
and use need an ethical guidance to avoid malicious uses and scandals like those revealed during
the last few years. Some of these events, like Cambridge Analytica, have had direct consequences on
democratic processes; other examples, such as the
use of facial recognition in mass surveillance and
law enforcement have had a significant impact on
our society, our daily living, and our relationship
of trust toward governments and new technologies.
As a consequence, in 2019 there were over 100
documents related to ethical principles published
around the world, including governmental agencies, but also big technological companies (e.g.,
Microsoft or Facebook), or academic and research
institutions, and the number is increasing. In [45],
authors analyze 84 of these AI guidelines and code
all the existing ethical concepts into 11 ethical principles. Among them are found the four principles
defined in traditional medical research: autonomy
(and freedom), nonmaleficence, beneficence,
and justice (and fairness). Other concepts can be
related to the Universal Declaration of Fundamental Rights and are strongly supported by the recently
published 2030 Sustainable Development Goals of
the United Nations [46], such as dignity, solidarity,
or sustainability. Finally, we can find principles that
are easily identified with dilemmas raised by the
use of AI: transparency, which is the most prevalent principle with 73 occurrences, responsibility,
privacy, and trust.
It is expected that public administrations that use
AI-based solutions act to create a positive impact
on society and environment, empowering citizens
with the use they do of the technology and offering
means to end users to be able to understand and
interact with it as well. One of the most critical issues
raised by AI is its impact on jobs and the economy
[4]. We should highlight that the responsibility to
build trust in AI relies on different stakeholders [47],
[48] and not only on the AI system. These stakeholders include different professional profiles, such
as IT/AI experts, lawyers and regulators, experts in

IEEE Technology and Society Magazine - March 2021

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