IEEE Technology and Society Magazine - June 2021 - 70

Feature
will always be the main guide. But there is also a
criticism of the modern, enlightened conception
of the values that permeate HR. HR puts too much
emphasis on the ethics of personhood or individual.
The philosophy of ubu-Ntu through its robust
conception of relationality invites us to reimagine
not only digital technologies, the Internet, and AI
but what it means to be human in an interconnected
relationship with the natural world and
other forms of life [13]. In the age of AI and robots,
the HR must take the central stage. Individual HR
such as the integrity of the person, autonomy,
self-determination, the right to privacy but also the
right to nondiscrimination, free speech, and the
right to political participation should guide the
development and implementation of AI technologies.
If directly or indirectly these HR are compromised
in the design or implementation of an AI
technology, this should cancel their deployment.
Human-centric AI and SDGs
Modern economy and society has evolved by
adopting technical advances such as machines for
automation of many tasks and activities. This automation
and deployment of machines including AI
technologies have helped society to try meet the
SDGs in a predictable manner. Society needs smart
and precise solutions and these are becoming available
through new applications of AI technologies.
Some examples of these applications of AI technologies
to meet the SDGs are
* sensors to measure CO2 emissions;
* smart farming: machine learning applications to
improve horticultural products;
* infection disease tracking and monitoring system;
and
* GPS and satellite imagery to monitor wildlife.
However, whatever applications are used to
meet the SDGs must be made from an ethical and
humane-centric perspective. With respect to the
ethical part, as we have commented above, this
must be done from a rigorous approach to ethics
and not from a simplified vision of it. Regarding
the humane dimension, every AI technology must
be applied in accordance to the HR framework.
One of the main axioms of human-centric AI is to
do everything possible to achieve value alignment:
how to align the values of highly capable intelligent
machines with those of humanity. Value alignment
70
problem consists in how to align AI with goals, values,
and preferences of its users when the potential
users can be all of humanity. A practical dimension
of the value alignment problem is to create artificial
systems that meet people's needs. AI applied
to SDGs is the corollary of the value alignment
problem. If we were able to harness the power of
AI to deliver on the SDGs, the problem of value
alignment could be largely solved and we will
have created a human-centered AI. Human-centric
AI means developing a new generation of AI
models that perform operations and tasks aimed at
improving the welfare of all, or at least, caring for
the planet, communities, and individuals.
Discussion
Ethics is an age-old discipline that should be
taken seriously when reflecting on the impact of
AI technologies. Most mathematical approaches
to several concepts within AI ethics (e.g., fairness)
do not take into account the complexity of what
" ethics " implies. Here, we wanted to emphasize
the idea that a principle-based approach to AI ethics
is not enough and that we must take HR as a
basic normative framework for AI technologies.
Our take home message is that to meet the SDGs
with the help of AI technologies, we must consider
ethics from the start (in the designing and building
of AI technologies). Another aspect to bear
in mind when dealing with AI ethics applied to
SDGS is acknowledging the mutual reinforcement
between three different components involved in
any algorithmic system:
Algorithms
↕
People
↕
Data
in this sense, the outcomes of an AI system will
depend on the mathematical and technical aspects
of the model, the data with which the model is
trained and the people who interact with the algorithmic
system. At any of these points in the process
there may be unintended consequences (e.g., bias)
and with a much greater impact, inherent complexity
or dimensionality when it comes to the deployment
of AI-powered systems in the context of SDGs,
so a rigorous ethical analysis of each of these components
is a necessity. One recommendation is to
IEEE Technology and Society Magazine
IEEE Technology and Society Magazine - June 2021

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