IEEE Technology and Society Magazine - December 2021 - 71

determine the impact that these councils will have
on shaping the global governance landscape for AI
and five other technologies, the initiative has the
potential to provide credibility and legitimacy to
these discussions and provide a powerful platform
to advance ideas.
Lessons may also be drawn from the International
Council for Harmonization of Technical Requirements
for Pharmaceutical for Human Use (ICH),
which was established in 1990 by the pharmaceutical
regulatory agencies in Japan, the United States, and
Europe, in partnership with their pharmaceutical
industry associations [27]. The resulting P3 provided
a platform to innovate within the field of pharmaceutical
science and harmonize their national regulatory
frameworks. The ICH strives to promote global
public health by encouraging common (yet not
identical) safety and performance standards across
jurisdictions [28], serving public and private interests
by speeding both market entrance and patient
access. By orchestrating multiple regulators with
expert industry input, the ICH has yielded " neither
a strengthening nor a weakening of national standards "
while rendering " national government regulation
more efficient and effective " [27].
These initiatives, along with others such as the
Global Fund to Fight AIDS, Tuberculosis and Malaria
(the Global Fund) and the Global Partnership for Sustainable
Development Data, highlight that success
may be achieved by grounding the development of
AI governance in a multistakeholder approach. We
argue in this article that a global P3 for AI, grounded
in multistakeholder approaches, could result in similar
benefits to national and supranational regulatory
efforts while inviting industry and civil society actors
to participate and contribute their technical expertise
to governance and harmonization projects.
Standards
Though at the " softer " end of the legal spectrum,
voluntary transnational instruments may also play
a significant role in the global governance of AI
[29]. These include technical standards issued by
standards-setting bodies such as the International
Organization for Standardization (ISO) or the IEEE
Standards Association (IEEE SA). Technical standards
can package regulatory norms as definitions
or as design or manufacturing specifications, which
can provide a common language or basis for measuring
the performance of a new technology. While
December 2021
compliance by firms is not legally required-unless
the standard is incorporated into domestic or international
law-it may become desirable to facilitate
international trade and reputational status in
markets impressed by standard compliance [30].
However, these benefits come primarily from global
standards, as multiple conflicting national standards
can complicate trade and regulatory orchestration.
Standards can be used to create and adjust regulatory
norms more quickly than through government
action, though with a process that may lack
transparency and democratic legitimacy. By operating
under consensus-based models where representatives
from around the world can participate
in deliberative norm creation, standards gain legitimacy
through their support from diverse groups and
stakeholders rather than from a legal mandate or
state sovereignty [31].
However, " pay for play " models which require
interested parties to pay fees to participate can jeopardize
this legitimacy by excluding public, private,
or civil society stakeholders without the means to
afford fees [32]. Kica and Bowman [32] point to
TC229-Nanotechnologies as a case-in-point. The
technical committee was established in 2005 in
response to the increasing public and private sector
investment in nanotechnologies and nanosciences,
and the increasing development of standards by
national and subnational bodies in response to
this activity. Despite significant investment in fundamental
research and development activities by
most states and Fortune 500 companies [33], and
the pressing need for standard harmonization, the
work of TC229 is being driven by-as of 2021-39
participating members, the largest of which is the
United States [34]. That is not to say that their work
output, which now includes 95 published ISO standards
as of November 2021, is not impressive and that
what they do is not of fundamental importance to
the development of the platform technology. But,
the drafting and establishment of these standards is
being undertaken by a limited number of players,
many of whom are key actors in shaping the governance
of nanotechnologies in other global forums.
Transnational standard-setting bodies have
already begun work on various technical aspects of
AI. In 2017, the ISO and International Electrotechnical
Commission's Joint Technical Committee 1 on
information technology formed a new subcommittee
42 (ISO/IEC JTC 1/SC 42) to craft and guide AI
71
IEEE Technology and Society Magazine - December 2021

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