IEEE Systems, Man and Cybernetics Magazine - July 2022 - 29

by Moloud Abdar, Abbas Khosravi, Sheikh Mohammed Shariful Islam,
U. Rajendra Acharya, and Athanasios V. Vasilakos
Introduction
In recent decades, an emerging phenomenon has entered
applied sciences, called AI. In a short time, the extraordinary
power of AI [i.e., machine learning (ML) and deep
learning (DL)] has solved a wide variety of problems in
many areas. Today, AI is used to solve complicated realworld
problems in various research domains. AI technologies
have been widely used in many critical systems,
such as autonomous vehicles (AVs), cybersecurity systems,
and many more [1], [2]. However, including different
AI techniques in such systems may pose risks [3]. It is not
feasible to devise AI-based systems, which know how to
appropriately operate in situations that differ outstandingly
from what AI systems have seen erstwhile. Consequently,
it is pivotal to develop the next generation of AI
systems that can perform adequately in known domains
and comprehend, demonstrate, and quantify when faced
with unknown situations. Among all AI-based technologies,
in particular, ML and DL methods have gained tremendous
attention in the mechanization of various
processes such as computer vision and image processing
[4]-[6], signal processing [7], gesture recognition [8], visual
question answering, text analysis and sentiment analysis
(SA) [9], zero-shot learning [10], medical data analysis
[11], big data analysis [12], and many more.
Today, advances in computer science and high-performance
computing find AI extensively benefiting the contemporary
world, such as with weather forecasting, fraud
detection, face recognition, and predicting diseases and
cancers [13]. AI methods have been applied to various
tasks, both in private and public sectors. The power of AI
has attracted considerable attention from researchers to
take advantage of its potential capabilities and subtypes.
The remarkable efficiency of AI has paved the way for
other branches of science and engineering to apply it in
their fields extensively. AI helps to build intelligent-based
machines that can improve various processes' performance.
The AI comprises a broad range of techniques
such as ML, DL, statistical methods, and expert systems,
which fundamentally depend on decision rules.
This means that only having an outstanding performance
of such intelligent methods is not a criterion for
selection or approval by medical experts. Comprehensive
reviews of previous studies indicate a close relationship
between trust and uncertainty, either in user
interfaces or information [14], [15]. In contrast, the
onset of new technologies such as AI, big data, ML, and
DL provide us with potentially significant insights into
a deeper understanding of raw data and the extraction
of meaningful and valuable information. Many people
still do not approve of these innovative technologies.
Recent studies have clearly shown that this lack of
trust stems from the weakness of these models in dealing
with their uncertainties [16], [17]. Trust is a concept
with different definitions. It should be noted that
some people claim that trust is only between people,
so we cannot " trust " machines of any kind. Humans
usually trust others who make fewer mistakes. But
trusting nonhumans is also a key challenge. But as
Smith and Hoffman [18] mentioned, trust can also be
designed for machines.
One of the most important applications of AI is
undoubtedly in medicine to diagnose diseases and cancers.
The rise of AI in medical data analysis has attracted
the attention of researchers worldwide. Physicians
have wanted to predict the disease early for years
before it becomes severe. Although attention by clinical
organizations is considered thus far, there is still a gap
between medical organizations and physicians with
automated systems. One of the main reasons for this
gap is the lack of trust by medical experts in either
July 2022 IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE 29
©SHUTTERSTOCK.COM/SCHARFSINN
http://www.SHUTTERSTOCK.COM/SCHARFSINN
IEEE Systems, Man and Cybernetics Magazine - July 2022

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - July 2022
