IEEE Consumer Electronics Magazine - November/December 2020 - 93

societies to what was supposed to be the merits of
technologies. The noise of global exchange of
information exposed the phenomena labeled as
fake news. Fake news has many forms and, before
the social networks technology, is mostly known
as "gossip". Yet, raised to a global level, it is a camouflaged threat that can influence the fate of
states.
Another very good example is the use of artificial intelligence (AI). AI rose from a well-coined
buzzword to one of the hallmarks of the second
decade of the 21st century.2 AI is a branch of computer science that aims to replicate or simulate
human intelligence in machines. The history of
modern AI begins in the mid of 20th century,
when Alan Turing explored the mathematical
possibility of an AI. Turing suggested that
humans use available information coupled with
reason to solve problems and make decisions, so
why cannot machines do the same thing? This
was the framework of his paper "Computing
Machinery and Intelligence" in which he discussed how to build intelligent machines and
how to test their intelligence.3 Six years later, the
first written AI program was presented at the
Dartmouth Conference 1956, when the term
"Artificial Intelligence" was ushered into community. Since then, AI has been experiencing highs
and lows, until last decade major breakthrough
in all segments of society via consumer technology. Coincidence? No. The AI success piggybacked on other technology developments that
enabled much needed computational power for
the deep learning models.
So where is the AI controversy? Ethics perhaps? Data collection and hidden understanding
of how AI reached the results may influence the
acceptance of the technology. In April 2016, the
European Parliament adopted the General Data
Protection Regulation that contains legal framework known as "right to an explanation." It emphasizes the principle of explainability for AI system's
decisions, "according to which it should be possible for them to be understood and traced by
humans." These AI guidelines provide focus for AI
designers in order to ensure fairness of the used
model and avoid unfair bias among the training
data. Furthermore, Filip Pie˛kniewski--a San Diegobased computer scientist and the author of an
essay titled "The AI Winter is well on its way"--

November/December 2020

said that much of the funding pouring into AI, particularly in the context of self-driving cars and
robotics, is based on unrealistic expectations of
what current deep learning is capable of doing. 4
Since AI models are driven by data, they are also
good as data they are being fed. As the old saying
goes: "Junk in - junk out." Or maybe a quantum
computing will fix this and turn a garbage into the
treasure?
Hard to comprehend the background of results
of AI methods, controversy of the 5G health
impact, fake news that spreads like a virus over
social platforms and communication applications
are just some examples that produce dents in
trusting the technology. This point to one important aspect of the technology development--security and protection. Many hours are spent in labs
to find ways to protect the user, to find business
models to protect data, or to secure technology
from malicious attacks, either by design or by
updates during implementation. But what about
vulnerabilities that are brute force?
We are witnessing frustration caused by lack of
adequately presented information that manifests
in demolition of 5G towers. Frequencies on which
a 5G network operates require a dense base stations infrastructure, approximately 100 m apart,
and each nonoperational tower can influence the
local operability of the network.5 However, is the
demolition of a physical infrastructure the only
security brute-force problem that can endanger a
dawn of the device-centric era? To make clear the
following example, let us analyze data. 5G highspeed Internet access has data throughput of
approximately 1 Gbit/s. Bandwidth around the
operating frequency is limited, and thus one of the
key issues with any modulation format is the spectral efficiency. Higher order modulation formats
are more spectral efficient, but also more vulnerable to interference. These formats are used when
signal to noise plus interference ratio (SNIR) is
high. As the communication link deteriorates,
order of a modulation format is reduced. Modulations of lower order provide lower throughput
within a certain bandwidth, but also are more
robust to interference. Hence, lower order modulation formats are used when SNIR is low. QPSK is
the lowest order modulation format used in 5G.
With QPSK modulation format, data throughput of
1 Gbit/s requires bandwidth of 1 GHz. When SNIR

IEEE Consumer Electronics Magazine - November/December 2020

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - November/December 2020