IEEE Technology and Society Magazine - December 2021 - 53

by private sector actors as well as governments to
encourage or discourage certain behaviors as illustrated
by the issuance of guidance documents and
voluntary calls for data [10]. These instruments are
created in a more agile way than traditional rulemaking,
can evolve and pivot in response to a changing
landscape, and are arguably less resource-dependent
than legislation or rules. As such, it has been
argued that soft law should play a central role in the
governance of emerging technologies [11], [12].
Soft law is not without its critics. Commentators
have noted that these mechanisms may lack legitimacy
and that the limited-if any-enforcement
provisions can result in varied and/or tokenistic
compliance, which can create an uneven playing
field for market participants [13]. However, such
mechanisms appear to be ideal for governing rapidly
evolving technologies, such as AI, due to the
ability to test different approaches across industries
and jurisdictions. Outcomes of this type of experimentation
can be utilized to better inform the design
and execution of subsequent instruments. It is also
important to note that soft law instruments do not
operate in a governance vacuum; rather, they exist
alongside legal frameworks and often work to supplement
the perceived deficiencies of hard law.
While often not thought of as a soft law mechanism
per se, the insurance and reinsurance sectors
have an important role to play in governing emerging
technologies such as AI. Between 2019 and 2021,
the Center for Law, Science, and Technology at the
College of Law at Arizona State University (ASU)
compiled a database of 634 current global soft law
programs focused on AI1 [14]. In this database, insurance
had no representation. This is, in many ways,
counterintuitive given the primary role of regulation-to
mitigate risk-combined with the fact that
the insurance sector is often the primary controller of
risk assessment and a key gatekeeper for accessing
the market. In this way, it easily fits with Marchant's
soft law definition, as it set substantive expectations
on companies who which to participate in the market
without governmental enforcement.
As noted by Trump et al. [15], " the role of insurance
as a risk management strategy has not changed
significantly in concept over several hundred years, "
with the calculation of risk being at the very heart
of their business. This holds true even when there is
1As of August 2021.
a high degree of uncertainty over risks and known-
unknowns. Given the ways in which the sector can
shape behavior through premium setting or denial of
coverage, this article argues that insurance will play
an important role in shaping the AI landscape and
rollout of certain applications. Some risks associated
with AI already fit easily into the tools offered by the
insurance sector; others, though, will need more
creative approaches. Nonetheless, the insurance
market may offer pathways for entities to innovate to
mitigate potential financial loss.
Nanotechnologies provide a case-in-point. As
detailed by Bowman, early concerns over the potential
risks posed by nanomaterials resulted in substantial
innovation across market sectors, organizations,
and jurisdictions in relation to the design and
implementation of soft law mechanisms [10] [16].
While their level of success varied, the true value of
these heterogeneous approaches is that they were
designed during periods of high scientific and societal
uncertainty and implemented in parallel with
the technology's development and commercialization-not
after the fact [10], [17], [18]. As highlighted
in the following section, the global insurance
sector was a pivotal player in the development of
this technology's governance.
Insurance as a soft law instrument for
nanotechnologies
The increasing commercialization of nanotechnology-based
products in the early 2000s occurred
against a backdrop of scientific uncertainty over
potential risks presented by nanomaterials to human
and environmental health and safety [19]. Within the
scientific community, little data existed in relation to,
for example, the potential toxicity of certain families
of nanomaterials, potential exposure pathways, and
the appropriateness of conventional risk assessment
protocols for determining toxicity of certain nanomaterials
[20]-[21]. Given the scale of the global investment
and the increasing integration of nanomaterials
into products across every sector, the technology presented
a complex and somewhat unique challenge to
the global insurance sector given that risk is a fundamental
variable used in calculating premiums.
It is, therefore, not surprising that some of the earliest
reports focused on how the technology should be
commercialized and governed were published by large
global insurance firms. In their 2002 report, Munich
Re noted the uncertainties posed by the technology,
December 2021
53
IEEE Technology and Society Magazine - December 2021

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