IEEE Technology and Society Magazine - December 2021 - 17

algorithms used to determine important aspects of
their lives [16].
While it is still too early to say for certain exactly
what went wrong, evaluating efforts against the
drivers of trust provides some insights. It appears
that a more inclusive, respectful approach to the
development of the whole initiative may have been
more effective. If underpinned by more humility in
the development and application of the algorithm,
a more respectful attitude to teacher assessments,
a more open approach that involved parents and
students, and better communication and explanations
for those affected and with all citizens from
start to finish of the process, disaster may have been
avoided. This is an illustration of why soft law that
seeks to better serve society, improve effectiveness,
and have democratic legitimacy should take steps to
involve stakeholders, including citizens, in defining
what the broader public interest is, what benefits to
people and planet are desirable, and what tradeoffs
are acceptable from the very beginning.
In his book Rebel Ideas, Matthew Syed's extensive
research makes a compelling case that " more of us
are better than some of us. " Diverse perspectives
highlight gaps in understanding, challenge assumptions
and biases, puncture information bubbles,
and reveal the limitations of echo chambers [17]. In
today's polarized and fragmented society, this may
be the only way to develop AI governance that will
serve the many and not a few.
Effectively understanding, synthesizing and adequately
reflecting the values and perspectives of so
many individuals is difficult, but it is not an insurmountable
challenge, as the recent OECD initiative
Innovative Citizen Participation and New Democratic
Institutions-Catching the Deliberative Wave [18]
has shown. Its almost 300 case studies about deliberative
processes demonstrate: " People are more
likely to trust a decision that has been influenced
by ordinary people than one made solely by government
or behind closed doors. "
Demonstrating that a soft law approach to AI governance
is in fact focused on the public interest-
particularly one which has actually been defined by
the public-may conceivably be the defining factor
for the legitimacy and trust in process and outcome.
The rest of this article will outline how designers of
new AI soft law ethical initiatives can go about defining
the public interest and earn the trust of citizens
and those participating in its governance-focused
December 2021
on three new capabilities of involving citizens, building
trusted environments, and communicating evidence
of trustworthiness.
Innovate by involving citizens
A common concern is that technology and governance
is far too complicated for ordinary people
[19]. But citizens are valuable when asked for
their knowledge, values, and expertise as citizens.
You do not need a PhD in AI to be able to engage
constructively on issues arising from the use of
facial recognition in policing for example, or your
views on the pros and cons of personal data use
in healthcare.
According to the OECD's Catching the Deliberative
Wave report, citizen involvement works best when
it is about " values-based decisions, those requiring
tradeoffs and areas demanding long-term solutions
beyond a single electoral cycle. " These are exactly
the types of issues presented by AI technologies and
which need the guiding compass that a focus on the
public interest provides.
Governance designers can be very skeptical
about the usefulness of involving citizens, but very
often citizens surprise them with their thoughtfulness,
the depth and nuance of their views, and the
important issues they raise which may not have been
considered [20].
Process design matters a lot in making the benefits
of citizen involvement happen-bad methods can
undermine trust in the process and the usefulness of
the outcome, while good ones can make it a fulfilling
and rewarding experience for everyone concerned.
Professor Beth Noveck, director of TheGovLab
at New York University, demonstrates in her book,
Smart Citizens, Smarter State, that if those designing
governance trust people to contribute their particular
knowledge, perspective, or expertise in a structured
way, they will not only empower them to become
more involved in their communities, but gain a huge
bank of expertise in the process [21], [22].
UNESCO is an example of an organization including
more voices and ideas into their process for governing
AI. As of July 2020, the organization launched
a global online consultation on the ethics of AI, to
give everyone around the world the opportunity to
participate in the work of its international group of
AI experts [23].
This illustrates that involving citizens need not,
and should not be a separate " public engagement "
17
IEEE Technology and Society Magazine - December 2021

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