IEEE Technology and Society Magazine - June 2020 - 53

T

his position paper develops an algorithmic approach to deliberation
as the basis for institutional reflexivity. We propose the idea of algorithmic reflexive governance for socio -techno - ecological systems,
bringing together interdisciplinary ideas from Artificial Intelligence and Democratic Theory in systems that recognize the interactions
between social, technological, and ecological actors and processes. We argue that this innovative
approach to conservation and sustainability can
make a significant contribution to mitigating the climate emergency.

Reflexivity
The dominant social, economic, and political institutions of the early 21st century seem to have difficulty
recognizing environmental preconditions and impacts,
and are either insensitive to the trajectory towards
catastrophe or lack the ability to adapt in a timely and
positive fashion. One diagnosis for this "institutional
sclerosis" is a lack of reflexivity [1].
Reflexivity, generally, is the ability of a structure, process, or organization to reconfigure itself in response to
reflection upon its own performance. The core aim of
this work is to develop self-governing socio-technical
systems where reflexivity with respect to ecological
and environmental impact is enabled through algorithms for deliberation, introspection and self-organization. To this end, we begin with a discussion of the
four dimensions of reflexivity, proceed to discuss an
algorithmic approach to deliberative processes that
resolve tensions in each of the four dimensions, and
finally use this as the recommended basis for developing socio-techno-ecological systems.

For each dimension of reflexivity (sources of knowledge, public discourse, institutional architecture and
institutional dynamics, there are two drivers (X versus Y)
each pushing the governance mechanism in a particular
direction. So, for sources of knowledge, there is one
push towards public participation (e.g., through the "wisdom" of crowds), an opposite push towards a technocracy. For composition of discourse, experiments in
opinion formation have shown how different parameters
of a social network can drive agents in one direction
towards "groupthink" (consensus on one opinion), or in
the opposite direction to diversity, in the range from
polarization (two opposite opinions) to fragmentation
(multiple opinions) [2]. For institutional architecture,
there is one driver toward centralization (one center of
decision making) or a client-server architecture, and an
opposite driver towards polycentricity (multiple centers
of decision-making [3]), i.e., a peer-peer architecture.
For institutional dynamics, in the study of cybernetics
and self-organizing systems, there is well-known "paradox" of trying to engineer stable systems equipped with
the capacity for adaptation, and yet adaptive systems
equipped with the means of preserving stability [4].
It is further argued [1] that deliberative processes
can handle the tensions between each of these binaries. In particular, deliberation can contribute to
ref lexive governance by managing these tensions
respectively: by joining expert and lay deliberation,
either face-to-face or by integrating the two in a broader
deliberative framework (e.g., citizen assemblies based
on sortition to provide political balance and inclusivity);
by generating meta-consensus, through validation of
multiple perspectives while establishing and maintaining practicable agreements between them; by enabling
deliberative learning through sharing information, experience, or best practice, and by transcending the binary

Four Dimensions of Reflexivity
In the context of environmental governance, it has
been argued [1] that there are four key dimensions of
reflexivity, each of which pits two plausible drivers in
tension (see Figure 1):
Sources of knowledge: public participation versus
expertise;
■ Composition of public discourse: diversity versus
consensus;
■ Institutional architecture: polycentricity versus centralization; and
■ Institutional dynamics: flexibility versus stability.

Expertise
Polycentricity

Stability

■

Diversity

Consensus

Flexibility

Centralization
Public Participation

Digital Object Identifier 10.1109/MTS.2020.2991500
Date of current version: 18 June 2020

JUNE 2020

FIGURE 1. The Four Dimensions of Reflexivity.

∕

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

53
IEEE Technology and Society Magazine - June 2020

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