IEEE Technology and Society Magazine - June 2020 - 65

global commons and the greater good. Tech giants'
algorithms are perpetuating discrimination [56] and
further dividing our social worlds [10], which corporations map and mine for profit: "Like the extraction
industries of previous centuries, they are highly
motivated to expand their territories and bend local
elites to their will. Without substantial pressure, they
have little incentive to serve a public beyond their
shareholders" [10, p. 32].
Public understanding is woefully limited of this use of
big data- and computation-based "narrow" AI in contemporary societies. Narrow AI is already shaping our
minds and practices every day when machines extend
human cognitive performance but in strictly delimited
tasks. It is built into our mobile phones, online search
engines, AI "helpers" such as Alexa and Siri, self-driving
cars, facial recognition, and bots that redirect our inquiries on webpages, leading us to additional products for
sale. AI algorithms thus guide us to what we learn and
sway our thinking, dispositions, and behavior, including
voting behavior. Our own micro-data, sold to the highest bidders in an emerging form of "surveillance
capitalism" [6], is the key to narrow AI's growing effectiveness. Yet, much journalistic and even academic
coverage of AI heralds the technology as a great boon
to public policy, especially in the environmental
space. Corporate media's cautionary tales about AI's
destructive potential focus more on general AI. The
recent movie Terminator: Dark Fate is one of the latest
Hollywood productions to do so, conveying fear and
human helplessness in the face of artificial intelligence gone haywire.
General AI is when the machines move beyond very
narrowly defined tasks to acquire a human level of intelligence applicable to any problem. They are the AI phenomenon on which futurists most fixate our fears, in the
form of terminators and other sentient machines boding
"dark fate" for humans. General AI is a relatively distant
reality, at best [57]. It is narrow AI that is shaping contemporary societies.
To the extent that the power of narrow AI is discussed, it is often portrayed as obscure and inscrutable
and as such elusive to human control [58, p. 3]. Descriptions of humans as "lacking adequate intellectual tools"
to engage the fast-changing, globally interactive actors,
technologies, and processes that shape algorithmic systems [59, p. 25] risk spreading mystifying perceptions of
algorithms as beyond human control. But AI is, by definition, controlled by humans. "Technologies can always be
understood at a higher level, intentionally [sic.] in terms
of their designs and operational goals and extensionally
in terms of their inputs, outputs and outcomes"; it is "[t]
raditional power structures [that] can and do turn systems into opaque black boxes" [60, p. 1].
JUNE 2020

∕

Current power structures also discourage full recognition of how technologies already are used to shape our
minds, to not call attention to the fact that they currently
do not sufficiently serve the common good. They limit
recognition that narrow AI can help us better grasp -
and therefore more consciously choose - the multi-scalar and diverse consequences of our choices in complex
systems, and to seize on its potential. This is "an historical opportunity to reset [societal] relationships in order
to distribute power and wealth more equitably" and to
foster societal transformations towards sustainability in
ways that are democratic, just, and that also reduce
socio-political divisions and suffering [61, p. 140]. Algorithms generating search results via Internet browsers
can be shifted away from current oppressive tendencies
[56] into the direction of knowledge and values conducive to these positive goals.
Transparency is essential for seizing on this potential: policies must force formalization and disclosure of
assumptions, choices, and adequacy determinations
associated with current and emerging automated and
intelligent systems [60].

Defining Alternatives
Responsible information and communications technology governance requires, and inherently involves,
normative and political decisions that shape the
future we get, because they exist to optimize whatever we tell them.
Therefore it is imperative to optimize them for the
world we actually want [61]-[63]. This agenda co-exists
uneasily with the cultural value scientists place on
value neutrality. Valorization of political neutrality
undermines the needed support for action-oriented
research on how to nurture the needed institutions and
transformations [64]. Besides being premised on an
elusive - and misguided - ideal, value neutrality in
research on institutions results in uncritical examination of existing societal arrangements [65] and integrates the questionable assumption [66] that value
neutrality is necessary to preserve scientific credibility
in socio-political arenas and that it is possible and
appropriate in media and technology policy [63]. The
(vitally needed) contribution of institutional critique and
proposals for reform are their subversive nature; the
fact of being subversive does not mean that they are
frivolous or lack rigor. Indeed, distance from value-neutrality increases responsibility for precise, accurate, and
well-grounded accounts [65].
Definition of what constitutes "good" alternative guiding principles and institutions requires assessment and
deliberation. It is not necessarily easy, but there are
sound starting points to be found. See, for example, [65],
[67]-[70] also, specifically, for governance of AI [5], [7],

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

IEEE Technology and Society Magazine - June 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - June 2020