IEEE Robotics & Automation Magazine - December 2021 - 121

regional, and/or national bodies
include the 2016 EU General Data Protection
Regulation, bilateral and multilateral
treaties, and the 2018 Council of
Europe " Modernised Convention for
the Protection of Individuals With
Regard to Automatic Processing of Personal
Data " (Convention 108+).
The IEEE Ethics Certification Program
for Autonomous and Intelligent
Systems [6] is a world first in setting
standards for the ethical certification
of products, services, and systems
deploying AI and robotics in the public
and private sectors. Certification is
essential to guarantee that these technologies
operate as expected when
they are interacting with human and
nonhuman agents. For stakeholders
involved directly and indirectly in the
lifecycle of AI and robotics systems,
certification guarantees that these systems
will cause no harm, their limitations
are known, and there will be
human accountability and responsibility
for their use. This, in turn, fosters
greater societal confidence in the technology's
utilization.
Different from these frameworks,
the standard developed by the IEEE
7007 WG has a formal and ontological
representation that can be used not only
as a foundation to elaborate public policies
but also to create computational
systems. In fact, IEEE Standard 7007 is
the first global ontological standard that
contains the concepts, definitions, and
axioms that are necessary to establish
ethical methodologies for the design,
development, and deployment of AI
and robotics.
IEEE 7007 WG
The IEEE 7007 WG is under the
umbrella of the IEEE SA P7000 series
devoted to ethics in A/IS. In this scope,
several WGs were formed-15 to date-
to deliver a broad range of standards
and/or recommended practices. Among
the goals of the IEEE 7007 WG are to
●
establish a set of definitions and
their relationships that will enable
the development of robotics and
automation systems in accordance
with worldwide ethics and moral
theories
●
align the ethics and engineering
communities to understand how
to pragmatically design and implement
these systems in unison
●
develop a precise communication
framework among global experts of
different domains, including robotics,
automation, and ethics.
To attain these goals, the IEEE 7007
WG developed a set of ontologies for
representing the domain in a more precise
way. As a result, IEEE Standard
7007 contains a set of ontologies that
represents norms and ethical principles
(NEP), data privacy and protection
(DPP), transparency and accountability,
and ethical violation management
(EVM). The 2development of this standard
was a complex process requiring a
dedicated lifecycle. For this purpose, the
IEEE 7007 WG developed an agile, collaborative,
and iterative methodology
called the robotic standard development
lifecycle [7].
The usefulness of ontologies in
standardization is twofold. On the one
hand, standardization processes are set
to produce a body of knowledge that
reflects a consensual view of practitioners
around a topic, defining, among
other aspects, a standard knowledge
structure in a domain, including common
concepts, relationships, and attributes.
Ontologies and their methods
provide a formal approach to that
aspect of the standardization process,
which is expected to produce a
sounder standard. On the other hand,
the ontologies themselves, as formal
artifacts, can be seen as products of
the standardization process that can
be used directly in data processing and
automatic reasoning. As an example,
one can cite IEEE 1872-2015 [8],
which set forth to establish clear definitions
for common terms in robotics
and automation.
IEEE 7007 Ontological Standard
for Ethically Driven Robotics and
Automation Systems
Top-Level Ontology
As a core ontology, the ethically driven
robotics and autonomous systems
(ERAS) ontology represents a midlevel
set of formalizations and commitments
that are platform independent and
intended to fit between an upper toplevel
or foundational ontology and
lower-domain and application-specific
ontologies. While some potential users
of the standard may intend to align the
ERAS core formalizations with existing
top-level ontologies specific to their
application domain, other user communities
will only require a minimal toplevel
set of conceptualizations to complete
the formalization of the concepts,
terms, and commitments axiomatized
in the ERAS ontology.
For that purpose, the four ERAS
subdomain ontologies are augmented
with axioms sufficient to complete
the definitions and commitments
expressed in the core ERAS models.
These axioms are expressed formally
using the Common Logic Interchange
Format (CLIF) [9]. The ERAS toplevel
ontology (ERAS-TLO) formalizations
define a minimal set of terms
deemed relevant to the characterization
of ethically oriented agents
and autonomous systems. It is not
intended to be applicable as a TLO in
other contexts.
NEP Ontology
The NEP ontology subdomain formalizes
the terminology and ontological
commitments associated with ethical
theories and principles that characterize
the norms of expected behaviors for
norm-oriented agents and autonomous
systems. This includes axioms for concepts,
such as norms, ethical theory, situation
plan repertoire, agent plans, plan
actions, and agent actions as well as the
corresponding relationships, such as
" selects plans from, " " subscribes to, "
" satisfies, " and " constrains plans for. "
Figure 1 depicts a brief and partial view
of a subset of the NEP terms with a
Unified Modeling Language (UML)
class diagram.
DPP Ontology
The DPP ontology represents concepts
and relationships among the diverse
agents, entities, and organizations that
may be involved at different stages in
data gathering, processing, transfer,
DECEMBER 2021 * IEEE ROBOTICS & AUTOMATION MAGAZINE *
121
IEEE Robotics & Automation Magazine - December 2021

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - December 2021
