Order this book
Carl Henning Reschke
University of Witten/Herdecke, Germany.
This is not the usual science or engineering book. Rather it is an insightful development and discussion of a vision and the beginnings of its implementation. A vision of how Artificial Intelligence and the Internet may evolve to become the electronic brain of human society, what this requires in technical terms and how it might ultimately be realised. This vision is quite well summarised by the term "wild computing" which refers to populations of agents that live and interact on the Internet, forming a large computer. The themes of the book link to issues in social ecology and evolution and to the emergence and development of intelligence. Some of the relevant topics have been discussed in popular books, for instance by Kevin Kelly in "Out of Control" (1994), by Howard Bloom in "The Global Brain" (2000) and by Maynard Smith and Szathmary (1995) from a biological point of view. Readers who have enjoyed these books and are interested in a treatment of ways to create evolving intelligent systems might also look at this book, even if they do not usually read "engineering" or the general systems literature. Since the book is largely non-formal and refreshingly well written, it is a quite easy read, and suffers only from the fact that some parts seem a little as if they were just "stuck together".
The goal of the book is to provide a framework for thinking about the Internet and what it could become. To do this, the book finds its starting point in philosophy and psychology and uses a complexity/general systems perspective. The book is structured in three parts with 15 chapters and an appendix giving definitions and more detailed explanations of some concepts. The first part discusses "Mind as Network", giving the "Practical Definition of Intelligence" used in the book, reviewing "Philosophy of Mind", discussing "Complexity and Emergence" and developing the "PsyNet Model of Mind."
The second part of the book deals with "Reconceptualising the Internet" as a networked evolving medium for "the" computer of the future. Here the computer is redefined as a network and a mind, and a "Fourfold Model of Internet Information Space" is developed. The "Internet Economy as Complex System" and the "Emerging Global Brain" are also discussed.
Part three ("Building Internet Minds and Worlds") describes the concept of a system of interacting agents. It is based on one language for knowledge representation and another for actions, both of which are described here. In "Webworld", a platform host for interacting agents is presented and its requirements are discussed from the perspective of Artificial Life. Then the book goes on to describe an Artificial Intelligence system for a "thinking" computer ("Webmind") that the author and his co-workers had started to develop in a commercial setting. Commercial application possibilities are also discussed in "Webmind on the Commercial Net". Taking a broader view, Goertzel gives his thoughts on social issues relating to the system in "Toward Superhuman Intelligence".
The appendices give a specification of the language KNOW, and present views and speculations on the autopoietic nature of life and evolutionary quantum computers related to Webmind and wild computing.
The review of themes and discussion of concepts is non-formal, concise and understandable even for uninitiated readers. It is a very good review in that it brings broad issues directly to the point and also manages to link them to each other. Probably the stance of psychological and complexity perspectives leads "naturally" to a certain tension with evolutionary principles. Despite many allusions to evolution in passing remarks, there is no explicit discussion of evolutionary issues. These might be important for the outcomes of interaction in a system of adaptive agents. For instance, in discussing emergence, Goertzel gives a very clear, not-too-technical formalisation that is much better than my attempt to come to terms with the phenomenon in this journal (Reschke 2001). It captures two types of emergence, which I have referred to as interactive and combinatory, but does not distinguish clearly between them.
From an engineering perspective, evolution might thus be considered a neglected issue within the framework used to implement the author's vision. We are dealing with the combination of network/software/hardware systems that are supposed to allow us to combine the capabilities of software applets and to merge software implemented "brains". Therefore, it would be interesting to know what happens when these are combined. The consequences of the ability to develop such hybrid systems would be far reaching. Research by Richard Watson and Jason Pollack at Brandeis University has led to the development of a simulation implementation which shows the effects of symbiotic combination mechanisms (without giving rise to emergence). Thus, from my point of view, the only weakness of the book is a lack of discussion of relevant evolutionary principles, which might impact on the design of the project and also be useful for its implementation.
The relevance of this book to social simulation is probably not so obvious from a description of the first two parts. On one hand, it lies in the discussion of principles necessary for the implementation of intelligent distributed systems. On the other hand it lies in the point of view of one exceptional thinker from whose reading one can go away inspired and with much food for thought. The third part deals with a conceptualisation of interacting agents in distributed systems, which is (of course) highly relevant. If the reader is interested in the material presented in the book, he may also find a look at the author's web site very interesting. Although the Webmind project had to be closed down, the issues raised by it are still being explored further. On the web site, there is more background material about the Webmind project and a lot of interesting information about other related subjects.
BLOOM H. 2000. Global Brain: The Evolution of Mass Mind from the Big Bang to the 21st Century. John Wiley, New York, NY.
KELLY K. 1994. Out of Control: The New Biology of Machines, Social Systems and the Economic World. Addison-Wesley, Reading, MA.
MARGULIS L. 1992. Symbiosis in Cell Evolution: Microbial Communities in the Archean and Proterozoic Eons, second edition. Freeman, New York, NY.
RESCHKE C. H. 2001. Evolutionary perspectives on simulations of social systems. Journal of Artificial Societies and Social Simulation, 4, <http://jasss.soc.surrey.ac.uk/4/4/8.html>.
MAYNARD SMITH J. and E. Szathmary 1992. The Major Transitions in Evolution. Freeman, Oxford.
Return to Contents of this issue
© Copyright Journal of Artificial Societies and Social Simulation, 2005