Kenneth Payne. I, Warbot: The Dawn of Artificially Intelligent Conflict. New York: Oxford University Press, 2021. 280 pp. $29.95 (cloth), ISBN 978-0-19-761169-2.
Reviewed by Andrea Rebora (King's College London)
Published on H-Sci-Med-Tech (October, 2022)
Commissioned by Penelope K. Hardy (University of Wisconsin-La Crosse)
Kenneth Payne’s I, Warbot: The Dawn of Artificially Intelligent Conflict is a compelling book exploring the complex intersection between artificial intelligence (AI) systems and warfighting. AI systems are already reshaping conflict, from autonomous drones and robot tanks deployed on the battlefield to data analysis leveraged by military officials and policymakers for decision-making support. They are changing how humans fight and will probably change the likelihood of the fights themselves. After explaining that popular understanding of AI is often based on science fiction, where robots look and behave like humans, Payne introduces his theory on warbots. Payne's title alludes to Isaac Asimov's short story collection I, Robot (1950). In "Runaround," a short story first published in Astounding Science Fiction in March 1942 and included in the 1950 collection, Asimov laid out his famous three laws of robotics, of which Payne presents a revised version: "A warbot should only kill those I want it to, and it should do so as humanely as possible"; "A warbot should understand my intentions and work creatively to achieve them"; and "A warbot should protect the humans on my side, sacrificing itself to do so—but not at the expense of the mission" (p. 12). These laws represent the pillars of Payne’s theory on warbots, such as the critical role of creativity in understanding human orders during wartime and the need for a pragmatic approach to killing during combat operations.
To make his work more accessible to an audience that might not be knowledgeable about the several aspects of AI, Payne describes how the research and development of AI systems had its ups and down in the second half of the twentieth century, how the increase of data available for training and computing power was critical, and how warbots can and will support military operations from the identification of targets to kinetic action against them. Historical guidelines for creating warbots do not exist, so Payne's solution is for potential warbot designers to understand how the desired warbot looks and behaves and to work backward. A central element of Payne’s idea for the development of effective warbots is the creation and use of systems that go beyond the highly effective but limited expertise required to dominate a single task, such as winning a game of chess or the board game Go, to the ability to make broad, strategic decisions.
To support his reasoning, Payne relies on the work of Prussian general and military theorist Carl von Clausewitz, in particular his recommendations for commanders in On War (1832). Clausewitz thought military commanders should possess "genius," the ability to bend fate to their will and seize opportunities others cannot see. As Payne explains, “genius required the ability to make decisions on the basis of imperfect information, in times of great stress, without being swept away by emotions.... It required more than a mechanistic following of rules or precepts.... His [Clausewitz’s] ideal commander was creative” (pp. 165-66). AI is excellent at processing data and identifying patterns, but it still lacks the creativity that is intrinsic to human nature. Warbots will be faster, more agile, and deadlier than today’s human-operated weapons, but the lack of general thinking and creativity makes them unable to cover the wide variety of scenarios that arise during warfare. To make up for it, and to leverage the advantages of both AI systems and humans, Payne suggests a collaboration between machines and humans where “the two very different forms of intelligence can combine to produce genuinely new and useful insights, covering for each other's respective weaknesses” (p. 182).
Finally, Payne’s research leads him to make three main predictions about the use of warbots in warfare. Despite the ethical issues of machines being capable of, and allowed to, decide who and when to kill, this technology is too valuable and advanced for countries to ban its research and development. Then, he explains how warbots are more effective when used offensively due to their swarming ability, that is, the ability to operate as a group with an interconnected network. Lastly, AI systems will excel at the tactical level in ways humans cannot match, but they will never be true strategists. “Even if machines make some, or many decisions about violence, war will remain something that is done by, and to, humans” (p. 84).
Despite the inherent challenges of researching and envisioning the development and application of emerging technologies, Payne provides a convincing look at what the future development and use of warbots might look like. His knowledge of history and warfare is highlighted in this work, and a couple of statements that might not be entirely correct from a technical point of view are overshadowed by his insights on war and conflict. Any reader, from undergraduate students and the general public to policymakers, can learn something valuable from Payne's work. In particular, students pursuing degrees in such fields as security policy and war studies will find many aspects to research and develop further in this book. Overall, it is a very approachable book that provides a good understanding of warbots and their challenges.
If there is additional discussion of this review, you may access it through the network, at: https://networks.h-net.org/h-sci-med-tech.
Citation:
Andrea Rebora. Review of Payne, Kenneth, I, Warbot: The Dawn of Artificially Intelligent Conflict.
H-Sci-Med-Tech, H-Net Reviews.
October, 2022.
URL: http://www.h-net.org/reviews/showrev.php?id=57745
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 United States License. |