Reviewed by
Camelia Florela Voinea
University of Bucharest
The volume Social Collective Intelligence. Combining the Powers of Humans and Machines to Build a Smarter Society edited by Daniele Miorandi, Vincenzo Maltese, Michael Rovatsos, Anton Nijholt and James Stewart provides a holistic approach of the social collective systems which embody various forms and technologies of social intelligence. Book’s chapters concerning currently investigated dimensions of social collective intelligence succeed to aggregate a comprehensive view of an interdisciplinary fast developing research area. It provides a state-of-the-art of the researches in domains like social collective systems, socio-informatics and social machines as well as in multiple computer science and artificial sciences areas concerning sensor-based information collection, processing, monitoring and management systems. Moreover, it analyses several major events from all over the world as relevant case studies aimed at illustrating the huge social impact of this kind of research, and its immense utility at all societal levels.The book consecrates a new view upon the role of machines in society as more and more dimensions of our physical and virtual world increasingly employ and (heavily) rely on machines, procedures and services which embody human intelligence. Innovation mainly based on the technologies of communication and artificial intelligence has induced a change in the way machines are perceived and understood. A new look on the society is currently emerging so as the classic industrial view of “man mastering machines” has turned into a view of “man cooperating with machines” which acknowledges the so-called hybrid systems. The concept of hybrid system is essentially based on the idea that no matter of its embodiment – biological or artificial, real or virtual – human intelligence makes humans and machines able to cooperate as autonomous goal-oriented agents in large-scale complex social systems.
In the Preface, the editors explain the motivation of their choices and give a brief presentation of the volume’s idea, of the domains covered, and of the major goals it serves. As it often happens, their book project started from an unavoidable research question: How to make the society a better place by viewing and constructing it as a hybrid environment? – The editors show that the book project made them realise that it would be the first which provides a holistic approach to the theoretical, technological and experimental research achievements in the targeted area. However, trying to find the answer to this question proved hardly challenging as the research area they had in mind was looking rather like a “puzzle”: fragmented and unevenly addressing various research issues due to the novelty of the research approaches and a relatively early stage of development. This guided the editors to conceive the book in a state-of-the-art perspective rather than attempting to achieve the clarity and completeness of a handbook:“[…] we aim to consolidate the fragmented research landscape, gathering contributions that capture the state of the art in all relevant areas, thus providing an up-to-date survey of existing research.” (Preface, p. iv). While concerned with the research and conceptual stage of such endeavor, the editors are also much concerned with giving: “[…] an overview of appropriate techniques that both scientists and practitioners can use in order to build purposeful and effective social collective intelligence systems.” (Preface, pp. iv-v).
The book is characterised by a high density of ideas and a wide range of research and application areas. The book’s structure reflects the richness of ideas, the complexity of the theme, and the diversity of contributions provided by multiple (teams of) authors.
The architecture of the volume comprises three basic modules which provide for a clear and self-explaining structure. Each of the three modules, Part I, Part II, and Part III, provides a condensed view over one of the three areas: (1) social collective intelligent systems, (2) technologies which have been embodied in the collective adaptive systems, and (3) applications explaining how the social collective intelligence actually performs, the benefits it brings, and the risks with which it is currently associated. Firstly, the volume introduces the basic concepts and theories concerning social collective intelligent systems, social informatics and social machines, and collective adaptive systems. Secondly, it further explains how concepts and theories relate and combine with technologies into social collective intelligent systems. Technologies for personal data monitoring and management are presented in order to explain their utility as well as the pitfalls of concentrating huge amounts of individual data on social platforms. Finally, it shows how this complex construction helps in developing applications targeting some of the common aspects of the daily life, like energy consumption, epidemiology and health care or humanitarian aid in crises and natural disasters.
Part I provides the fundamental issues of the social collective intelligence concept and use, covering the philosophy of the concept, its major aims, the most important aspects of moral and pragmatic justification, and the research means and goals in which it has found support.
In Chapter 1, Toward the Ethical Governance of Smart Society, Mark Hartswood, Barbara Grimpe, Marina Jirotka and Stuart Anderson provide a comprehensive approach on the ethical governance of a Smart Society. The concept of Smart Society has been elaborated as a complex extension of the Smart City idea. However, scaling up the Smart City architecture, mainly based on the idea of augmentation of social services by employing intelligent systems, to the Smart Society architecture, mainly based on the idea of global governance of a functional structure of (large-scale evolving) Hybrid and Diversity Aware – Collective Adaptive Systems (HDA-CASs) proved to be a challenging endeavour. The chapter introduces the basic notions of hybrid computing systems which define how joint work of humans and machines could provide for the emergence of new problem-solving capabilities. As a difference from other approaches on collective adaptive systems (Loreti and Hillston, 2016), the authors focus on HDA-CAS systems as involving both human agency and machines in a collaborative, interdependent framework able to tackle conflicting co-existent perspectives and preserve social values. Functional structures of such systems require the approach of the identity of the human actors, their skills, tasks complementarity and action coordination along with the preservation of social values and norms. As the collective adaptive systems prove a high degree of unpredictability due to emergent properties, their benefits are counterbalanced by risks. Ethical governance of social collective intelligent systems thus requires mechanisms which provide for preserving the human and social values while allowing the diversity and co-existence of conflicting issues. The main contribution of this chapter focuses on layered and intersecting principles of ethical governance based on specific operational paradigms: polycentric, hierarchical, environmentally embedded, adaptive and motivational. It explains how these can be combined in HDA-CAS mechanism architectures designed as value sensitive adaptive constructs: the governance design of a Smart Society is based on combining knowledge about collectives and of their social values and modes of regulation with knowledge and information provided by sensitive structures in HDA-CASs. The case study describes and explains the conceptual model for a specific scenario of home care. The focus on governance design allows the authors to introduce the design paradigm of Responsible Research and Innovation (RRI) employing principles of social embeddedness and responsiveness of innovation.
The approach developed by Jeremy Pitt, Didac Busquets, Alkaterini Bourazeri and Patricio Petruzzi in Chapter 2, Collective Intelligence and Algorithmic Governance of Socio-Technical Systems concerns the issues of self-governing socio-technical systems, self-organisation and resource management in collective intelligent systems. Self-governing institutions (open systems) need a paradigm for common-pool resource management. The contribution of this chapter consists in elaborating an algorithmic basis for self-organising resource allocation systems embedded in social collective intelligent systems (pp.32). The authors show that both natural and retributive justice principles (Ostrom, 1990) and distributive justice principles (Rescher, 1966) can be axiomatised in computational logic and made available to users in the form of a computational justice system. The system ensures “fairness in resource distribution over time” (p.35). It is based on two mechanisms inspired by Ostrom’s works, collective awareness and social capital, which provide for the management of community resources: while collective awareness, a mechanism based on social network communication, provides support for action coordination, the social capital mechanism is employed in the self-organisation of users’ flexible demands as an enhanced collective problem-solving capability. The case study presented in this chapter concerns a common resource (energy system) which is used and managed by a community of users.
Paul Smart, Elena Simperl and Nigel Shadbolt provide in Chapter 3 a taxonomic approach on social machines. The concept of social machine represents a complex generalisation of the notion of social collective systems. The authors justify their approach by the need to understand, identify, evaluate and properly classify hybrid systems in which humans and machines contribute with different skills and tasks by goal-directed participation and/or competence-based and task-based interaction. Earlier research literature on this subject defines, for example, crowdsourcing (introduced by Howe in 2006; see also an excellent literature review in Hossain and Kauranen, 2015) and collective intelligence (Bonabeau, 2009) as such social machineries. This approach aims at identifying a characterisation (that is, a set of attributes) which could be general enough such that social machines can be properly defined and efficiently designed. The authors define a social machine space (morphospace), a biologically-inspired notion of taxonomic space. They make an important step forward in achieving definitional generality as compared with the social machine concept introduced in earlier approaches: while remaining at the web-based view on social machines: “Social machines are web-based socio-technical systems in which the human and technological elements play the role of participant machinery with respect to the mechanistic realization of system-level processes” (Smart and Shabolt , 2014, p. 54), the authors make explicit two main concerns in properly classifying a participant machinery as a social machine. For one regards the capacity of social machines to support extended capabilities of joint work of humans and machines which cannot be identified on a complementary basis in either of them, nor can they be identified in less integrated systems (p. 56). The other one concerns the processes that involve the joint contribution of multiple individuals (humans) and one (or more) web technology based machineries: “[…] we discern a social machine when we encounter a process that demands a (mechanistically-oriented) explanatory account formulated in terms of the joint contributions of multiple individuals and Web-based technological components. […] social machines are the physical systems that perform, implement or realize such processes.” (ibid., pp. 54-59). Compared with the definition of social machine as a web-based machine (Berners-Lee and Fischetti, 1999), this definition extends the social machine morphospace so as to cover both extant and potential machineries, both web technology-based social machines and machines based on other technologies still to be discovered. Besides this, it classifies more accurately extant social machines by carefully identifying the type and proportion of participation and interaction between humans and machines. To the aim of understanding the relevance of these concerns in achieving the taxonomy of social machines, the authors provide an exquisitely elaborated appendix including a comparative analysis of these aspects in a systematic manner.
Part II focuses on the main technologies of data collection, storing, automated processing and use in social collective intelligence systems (SCIS) and collective adaptive systems (CAS). The three chapters in this part analyse both user and service providers’ concepts of data access rights management.
In Chapter 6, Privacy in Social Collective Intelligence Systems Simone Fisher-Hubner and Leonardo Martucci concerns the technologies employed in privacy protection and the risks induced, especially by the socialising network, in terms of reputation, trust, profiling and provenance in SCIS. The authors base their approach on the legal provisions of EU and OECD legal frameworks concerning the right to privacy and make a thorough evaluation of the actual and potential means employed by the web and information processing technologies for protecting the right to privacy of persons and collectives. The approach is valuable as it contributes substantially in defining and explaining the principles and methodology of privacy-enhanced social collective intelligence systems which balance privacy with transparency and security goals and techniques.
In Chapter 7, The Future of Social Is Personal: The Potential of the Personal Data Store, Max Van Kleek and Kieron Ohara approach the concept of management and the relationship between what individuals and collectives want to do with their personal or group data, on the one hand, and what the service providers actually do with this big data. The authors show that what a social collective intelligence system can achieve is a proper understanding and a better approach of personal data access and distribution management informed by the theoretical domain of Personal Information Management. They advocate the idea that the direct reach to personal data on the Web should be balanced by a concept of management able to regulate the ways in which a small number of providers succeed to centralise transfer of access rights to third parties. In the sensitive context of expanding technologies supporting the propagation of fake news on the Web, the centralising tendency make the Web ecosystem ”increasingly fragile” by determining “a majority of Web users … [to] rely on an oligarchy of service platforms which are in turn amassing a disproportionate quantity of users; personal information” (p.127).
The issues of reputation, trust and provenance are approached in Chapter 8, An Auditable Reputation Service for Collective Adaptive Systems by Heather Packer, Laura Dragan and Luc Moreau. The authors concentrate on collective adaptive systems: multiple interacting agents can trust data and service providers by relying on provenance data and auditable reputation services, which aim at increasing public awareness. The employment of these mechanisms could help in promoting accountability principles in the management of data access rights and in building providers’ reputation (pp. 163-164).
Part III of the volume consists of five chapters. The chapters present applications and case studies in the areas of online education platforms (Chapter 9), computational epidemiology (Chapter 10), urban hybrid systems (Chapter 11), and social collective intelligence systems in crises and humanitarian aid (Chapter 12).
In Chapter 9, Surfacing Collective Intelligence with Implications for Interface Design in Massive Open Online Courses by Anna Zawilska, Marina Jirotka and Mark Hartswood, the authors provide a foundational work aimed at providing a better understanding of the massive open online courses (MOOC) and supporting its design and development as a SCI system. The chapter addresses the educational platforms as open systems with the help of which students can ‘make relations’, that is, they can create experience-based links between the real world and the knowledge they acquire while using the platform.
Magnus Boman approaches in Chapter 10, Who Where When? On the Use of Social Collective Intelligence in Computational Epidemiology the use of social collective intelligence systems as support for developing computational epidemiology. The approach shows that SCIS could provide support to the development of new data analysis methods and tools which could offer the chance to employ big data in due time so as to support health care programs at societal level.
Chapter 11, Social Collective Awareness in Socio-Technical Urban Superorganisms by Nicola Bicocchi, Alket Cecaj, Damiano Fontana, Marco Mamei, Andreea Sassi and Franco Zambonelli approaches the issue of transformation of urban environments into socio-technical urban systems called ‘urban superorganisms’. In such environments, people are connected with each other and with the intelligent systems in ways which make them aware (collective awareness) of the urban context by means of sensors networks and computational resources included in the social networking infrastructure (pp. 236-237). This kind of infrastructure would provide for advanced urban services covering everything from intelligent transportation system to environmental sustainability and participatory governance.
Chapter 12, Collective Intelligence in Crises by Monika Buscher, Michael Liegl and Vanessa Thomas presents the role played by collective intelligent systems in response to natural disasters. Three case studies – Haiti earthquake, flooding in Alberta, Canada, and Boston Marathon bombing – provide a detailed picture of the ways in which socio-technical systems could and should be employed in providing humanitarian aid and re-organize community life during and after crises, natural disasters or terrorist attacks.
The volume provides a comprehensive view on the design of social collective intelligence systems, on innovation and most relevant research questions: Chapter 5, Twelve Big Questions for Research on Collective Intelligence by Stuart Anderson, Daniele Miorandi, Iacopo Carreras and Dave Robertson, and Chapter 13, The Lean Research: How to Design and Execute Social Collective Intelligence Research and Innovative Projects by Daniele Miorandi, Iacopo Carreras and Imrich Chlamtac.
It also offers a view upon the changes in the professional mentalities of scientists and researchers in the context of SCIS development: Chapter 4, The Mathematician and the Social Computer: A Look into the Future, by Martin Golumbic.
The volume aims at endorsing a style of research development and a research culture oriented towards making the society smarter and safer. It also endorses a culture of responsible socialising and responsible media use, solidarity and compassion for human condition in crisis situations. It is this particularity of the approach which makes this book worth reading: on the one hand, it offers knowledge, information and high-level expertise in a new area of study widely open for students in social and political sciences as well as in computer sciences and the sciences of the artificial: artificial intelligence, artificial life, machine learning, artificial autonomous agents, agent-based systems or complex adaptive systems. On the other hand, the volume is conceived as a “map” of a new area of scientific investigation on which the authors have marked the strategic points of interest, strategic local and global resources, the open issues and (still) unanswered research questions as well as the actual or potential pitfalls, risks and damages. It thus equally addresses and makes interested students, researchers and academics, companies, administrations and institutions.
ReferencesBONABEAU, E. (2009). Decisions 2.0: The Power of Collective Intelligence, MIT Sloan Management Review 50(2), pp. 45-52.
HOSSAIN, Mokter and Kauranen, Ilkka (2015). Crowdsourcing: A Systematic Literature Review (May 4, 2015). Available at SSRN: https://ssrn.com/abstract=2602176.
HOWE, J. (2006). The Rise of Crowdsourcing, Wired, Vol. 14 No. 6, pp. 1-4.
LORETI, M., Hillston J. (2016). Modelling and Analysis of Collective Adaptive Systems with CARMA and its Tools, in: Bernardo M., De Nicola R., Hillston J. (Eds.): Formal Methods for the Quantitative Evaluation of Collective Adaptive Systems. SFM 2016. Lecture Notes in Computer Science, Vol. 9700, Springer.
OSTROM, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
RESCHER, N. (1966). Distributive Justice. Bobs-Merrill, Indianapolis.
SMART, P.R., Shadbolt, N.R. (2014). Social machines. in: Khosrow-Pour, M. (Ed.), Encyclopedia of Information Science and Technology, IGI Global, Hershey.
Return to Contents
of this issue