Skip to main content
Top

2018 | Book

Carl von Clausewitz, the Fog-of-War, and the AI Revolution

The Real World Is Not A Game Of Go

insite
SEARCH

About this book

The language of business is the language of dreams, but the language of war is the language of nightmare made real. Yet business dreams of driverless cars on intelligent roads, and of other real-time critical systems under the control of algorithmic entities, have much of war about them. Such systems, including military institutions at the tactical, operational and strategic scales, act on rapidly-shifting roadway topologies whose ‘traffic rules’ can rapidly change. War is never without both casualty and collateral damage, and realtime critical systems of any nature will inevitably partake of fog-of-war and frictional challenges almost exactly similar to those that have made warfare intractable for modern states. Into the world of Carl von Clausewitz, John Boyd, Mao Tse-Tung, Vo Nguyen Giap and Genghis Khan, come the brash, bright-eyed techies of Alphabet, Microsoft, Amazon, and Uber who forthrightly step in where a phalanx of angels has not feared to tread, but treaded badly indeed. In this book we use cutting-edge tools from information and control theories to examine canonical and idiosyncratic failure modes of real-time cognitive systems facing fog-of-war and frictional constraints. In sum, nobody ever navigates, or can navigate, the landscapes of Carl von Clausewitz unscathed.

Table of Contents

Frontmatter
Chapter 1. AI in the Real World
Abstract
Straightforward arguments from control and information theories imply that emergence of the AI revolution from games of Chess and Go into the real world will fatally encounter the central matters of Carl von Clausewitz’ analysis of Zweikampf warfare. Promises of graceful degradation under stress for large numbers of driverless vehicles on intelligent roads, of precision targeting that avoids civilian collateral damage for autonomous or so-called man/machine centaur weapons, of precision medicine under even normal living condition, let alone during the current slow disasters of climate change and social decay, of the ability to manage financial crises in real time with agent-based models, and so on, are delusive groupthink or marketing hype that will be beta-tested on human populations, a gross contravention of fundamental moral and legal norms.
Rodrick Wallace
Chapter 2. Extending the Model
Abstract
It is possible to extend the model to nonergodic cognitive systems, a parallel to the nonparametric extension of more familiar statistical models. Cognition of any nature involves choice that reduces uncertainty. Reduction of uncertainty implies the existence of an information source dual to the cognitive process under study. Information source uncertainty for path-dependent nonergodic systems cannot be described as a conventional Shannon entropy since time averages are not ensemble averages. The fact that information as a form of free energy, however, allows study of nonergodic cognitive systems having complex dynamic topologies whose algebraic expression is in terms of directed homotopy groupoids rather than groups. This permits a significant extension of the Data Rate Theorem linking control and information theories via an analog to the spontaneous symmetry breaking arguments fundamental to modern physics.
Rodrick Wallace
Chapter 3. An Example: Passenger Crowding Instabilities of V2I Public Transit Systems
Abstract
We apply the theory to passenger crowding on vehicle-to-infrastructure (V2I) public transit systems in which buses or subways become so crowded that they are ordered by a central control to begin a degraded ‘skip-stop’ service. Application of the Data Rate Theorem shows there is no coding or other strategy that can compensate for inadequate service levels that produce passenger crowding of either stops or vehicles.
Rodrick Wallace
Chapter 4. An Example: Fighting the Last War
Abstract
We examine how inadequate crosstalk between ‘tactical’ and ‘strategic’ levels of organization will lead to another version of the John Boyd mechanism of command failure: the rules of the game change faster than executive systems can respond. Adequate levels of crosstalk take work.
Rodrick Wallace
Chapter 5. Coming Full Circle: Autonomous Weapons
Abstract
The powerful asymptotic limit theorems of control and information theories illuminate target discrimination failures afflicting autonomous weapon, man/machine centaur or cockpit, and more traditional structures under increasing fog-of-war and friction burdens. Degradation in targeting precision by high level cognitive entities under escalating uncertainty, operational difficulty, attrition, and real-time demands, will almost always involve sudden collapse to the familiar pathological ground state in which all possible targets are enemies, historically known as ‘kill everyone and let God sort them out’.
Rodrick Wallace
Chapter 6. An Evolutionary Approach to Real-Time Conflict: Beware the ‘Language that Speaks Itself’
Abstract
We examine real-time critical processes through an evolutionary lens, finding that protracted conflict between cognitive entities can trigger a self-referential, coevolutionary bootstrap dynamic, virtually a ‘language that speaks itself’. Such phenomena do not permit simple command-loop interventions in John Boyd’s sense and are very hard to contain.
Rodrick Wallace
Chapter 7. Summary
Abstract
The language of business is the language of dreams, but the language of war is the language of nightmare made real. Yet business dreams of driverless cars on intelligent roads, and of other real-time critical systems under the control of algorithmic entities, have much of war about them. Critical real-time systems, including military institutions at the tactical, operational and strategic scales, act on rapidly-shifting roadway topologies whose ‘traffic rules’ can themselves rapidly change.
Rodrick Wallace
Backmatter
Metadata
Title
Carl von Clausewitz, the Fog-of-War, and the AI Revolution
Author
Rodrick Wallace
Copyright Year
2018
Electronic ISBN
978-3-319-74633-3
Print ISBN
978-3-319-74632-6
DOI
https://doi.org/10.1007/978-3-319-74633-3

Premium Partner