IEEE Technology and Society Magazine - December 2021 - 68

Soft Law Governance of Artificial Intelligence
A Survey of Instruments
and Institutions
Available for the Global
Governance of Artificial
Intelligence
Walter G. Johnson
School of Regulation and Global Governance
(RegNet), Australian National University, Acton,
ACT 2601, Australia
 EmErging tEchnologiEs havE always challenged
traditional regulatory regimes [1]. This is
not new nor is this tension likely to change anytime
soon given the increasing speed at which embryonic
technologies are emerging into the market [2].
Some of these technologies, such as autonomous
vehicles or drones, can be effectively regulated at
the national level due to clearly defined jurisdictional
boundaries and the existence of relevant
national and state/provincial regulatory agencies
that oversee their conventional counterparts. While
the entry of these technologies into the market will
challenge regulators and policymakers and may
require tweaking of existing regimes, we argue that
these technologies, and the products that they enable,
will not require a sui generis response at the
national and supranational levels. The same cannot
be said for artificial intelligence (AI) and the myriad
of interconnected and interwoven applications
made possible by AI.
As evidenced by technological unemployment
from automation and algorithmic bias in the public
and private sectors, AI is already challenging existing
Diana M. Bowman
Sandra Day O'Connor College of Law, Phoenix,
AZ 85004 USA
Digital Object Identifier 10.1109/MTS.2021.3123745
Date of current version: 7 December 2021.
68
0278-0097/21©2021 IEEE
national legislative regimes [3]. Moreover, with existing
AI applications showing little respect for jurisdictional
boundaries, the ability to enforce legislation
is limited. We argue that this fissure will only be
widened if AI regulation is left solely in the hands of
national and/or supranational governments.
This article argues that effective governance of
AI will require international coordination and cooperation,
underpinned by a flexible and dynamic
governance framework. As this article seeks to highlight,
there are multiple approaches that could be
employed for AI ranging from hard law to various
forms of soft law [4]. Based on our analysis of these
governance frameworks, we argue that the most
robust framework for AI is one that adopts a multistakeholder
governance approach: one that is able
to continually adapt to emerging AI systems and
threats. Moreover, we argue that the development
of this governance approach must begin now-not
tomorrow, not next year-so as to allow for the
necessary iterations that this ubiquitous technology
demands. While we do not suggest that the development
of such a framework will be simple, given
today's political realities, we argue that such a multistakeholder
governance approach can be realized
if the necessary leadership from across sectors and
domains work collectively to do so.
IEEE Technology and Society Magazine
IEEE Technology and Society Magazine - December 2021

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