Page 49 - ITU Kaleidoscope 2016

HUMAN SAFETY CONSIDERATIONS IN THE EMERGING ICT ENVIRONMENT
Shailendra K. Hajela

ABSTRACT knowledge of the results of such action. In a situation, a
robot is flooded with a constant stream of perceptual data
This paper deals with the challenges posed to human safety that can be incomplete, and information and insights
in the emerging IoT and ubiquitous network scenario by reached in a fuzzy way. In order to use this information
super intelligent machines/ cognitive robotics that would effectively for determining actions, a robot must have the
coexist in the upcoming heterogeneous world comprising ability to categorize its experience. Based on these
humans and super intelligent machines. Questions of trust categories, a robot must be able to predict how the
in the super intelligent machines have been raised by environment will change as a result of its actions. Most
scientists, sociologists, industry stalwarts, etc. After importantly, this process of development should be driven
deliberating on the issues, aligning AI to human interests by an intrinsic motivation to explore the categories of its
has been suggested as a plausible option. Agenda for experience where it can make the most learning progress.
further action proposed is that the ITU may constitute a
multidisciplinary umbrella group, collaborating with The ITU in its report [2] on IoT (2005) identified the field
existing technical and policy groups working on AI of robotics as one of the starting points for the creation of a
standards at technology and policy levels, for in-depth world of smart things, which has since experienced steady
study of all facets, bridging the technology innovations and growth. During the past decade, the most important factors
the policy considerations, and bring out Guidelines for AI affecting the rate of growth of the robotics market have
aligning with human interests adhering to ethical code. been their enhanced capabilities and functionality. The
market size of industrial robotics initially has been greater
Keywords— Human safety, IoT, Trusting Robots, than the market size of personal and service robotics.
Cognitive informatics, AI Rules for super-intelligent However, according to the report, the personal robotics
machines segment is likely to lead future market growth. This would
include a wide range of different kinds of robots to do
1. INTRODUCTION house chores, industrial jobs, and personal care giver
partner-robots. In Japan, one of the main drivers of the
Rapid advancements in the fields of Internet of Things, personal robotics market is a rapidly aging population – by
artificial intelligence, super smart machines, cognitive 2050, more than a third of all the population is estimated to
informatics, cognitive robotics, soft computing, and related be 65 or over in age, creating a lucrative market potential
areas have aroused interest as well as concerns about the for the elderly-care robots, according to the report. Super
emerging interactive man-machine heterogeneous smart Robots that can also express emotions are already in
environment, where autonomous super intelligent the market place.
machines/robot working solo or teaming with, can match
and even surpass humans in speed and precision. No longer It may be noted that in this paper, Artificial Intelligence
is this a matter of science fiction when Isaac Asimov in (AI) is assumed to be embedded or otherwise leveraged in
1942 made the 3 (technically 4) laws to assign paramount the super smart machines and robots.
importance to human safety [1]:
1. A robot may not injure a human being or, through Some years ago, Google reported [3] developing an
inaction, allow a human being to come to harm; intelligent machine, an autonomous self-driving vehicle
2. A robot must obey the orders given it by human navigating itself through other vehicular traffic and hazards.
beings, except where such orders would conflict with The resulting boon to society due to its artificial
the First Law; intelligence that would ensure adherence to traffic rules and
3. A robot must protect its own existence as long as such impart faster navigational and decision making capability to
protection does not conflict with the First or Second avoid accidents would be enormous. It would eliminate
Laws. driving fatigue, instant human judgment and navigational
0. A robot may not injure humanity, or, by inaction, allow errors and would make the commute safer to arrive
humanity to come to harm. precisely at the set out destination. While one could feel
Adaptations and extensions exist based upon this happy for the truck drivers who drive long hours to ferry
framework. But these have for long remained in the realm goods thousands of kilometers, one would be concerned
of science fiction. The developments in the field of AI, with its turning into a killer-on-wheels should an
Robotics, Cognitive Informatics, IoT (Internet of Things), inadvertent malfunction occur or due to some rogue person
and related fields warrant their no longer being considered managing to successfully intrude and take over its control
fictional. A super intelligent machine, say, a cognitive robot over the Internet. Not to forget that the truck drivers role
is designed to perceive, analyze, reason out and decide its could change or even be eliminated. With the IoT, it also
action based on its past experience and predictive gets its identity established and can be accessed. Due to

978-92-61-20431-0/CFP1668P-ART © 2016 ITU – 31 – Kaleidoscope