- Book: "The Age of AI: And Our Human Future" by Henry A. Kissinger; Eric Schmidt; Daniel Huttenlocher. – New York: Little, Brown and Company, 2021. ISBN 978-0-31627380-0; 978-0-31639441-3
- Review by Isak Runarsson
“Each major technologically advanced country needs to understand that it is on the threshold of a strategic transformation as consequential as the advent of nuclear weapons – but with effects that will be more diverse, diffuse, and unpredictable.” Such is the prognosis of a star-studded line-up of authors in The Age of AI And Our Human Future (2021). The authors compellingly argue that the role of Artificial Intelligence (AI) in human societies is set to become much more important and even dominant in everything from the mundane decisions of everyday life to the grand strategic questions of statecraft.
In a larger context, the book serves as an alarm bell. At the onset, the authors tell a story of two of them meeting at a conference where one of them intended to skip out on a session on AI, assuming that the discussion would be technical and not to his interest. At the urging of one of his later co-authors he ended up going to the session, a decision that eventually sparked the writing of their book. The story can be generalized in that policymakers, lawyers and statesmen often have limited enthusiasm for technical and technological issues, which can lead to suboptimal levels of knowledge about potential disruptive impacts of said technology. The reverse, that engineers and technologists do not consider thoroughly enough the many societal implications of their work, may also be true. Such may be the fate of artificial intelligence now.
The sheer (human) intellectual firepower devoted to the effort of explaining AI and its implications can in and of itself be seen as a cry for a more dynamic public discourse on the subject and a call to action. The former U.S. Secretary of State and master statesman Henry A. Kissinger needs no introduction, but his co-authors are no less leaders in their fields of expertise. A PhD from University of California Berkeley, Eric Schmidt served as the CEO and Chairman of Google for a decade and as chairman of Google’s parent company Alphabet for seven years. He is a leading technologist and expert on the business and economic implications of technological development. The second of Secretary Kissinger’s co-authors is Daniel Huttenlocher, the Dean of MIT’s Schwarzman College of computing and Professor of Computer Science and AI & Decision Making at the same school. Huttenlocher is among the leading experts in academia on Artificial Intelligence.
In order to increase mass appeal, the authors have somewhat had to sacrifice clear organization of the book’s contents. Overall however, the first four chapters seek to explain the current capabilities and limits of AI, how it differs from human intellect, and where and how it is currently being operated. The first chapter of the book begins by explaining the current capabilities of AI. This is primarily accomplished by using a board game setting as an example, a favoured analogy of Kissinger’s (in his prior book On China, he sought to explain the differences in strategic thought between China and the West through each civilization’s choice of board games). The second chapter takes the reader through a whirlwind history of thought on human reason, intelligence and conceptions of reality to set up the argument that AI represents a new paradigm in our collective formulation of these issues. The third chapter is a logical extension of the first chapter, diving deeper into the different types of AI and their properties. The fourth chapter then presents a case study of how prevalent AI is in our daily lives without many of us knowing it – the extensive use of AI in network communication platforms often goes unnoticed. However, the title Global Network Platforms and setup of the chapter can be a little confusing, as it can at times seem like the authors are implying that AI is part and parcel of network platforms even though they can be operated with and without AI. The same problem also at times surrounds the book’s discussion of cyberwarfare, although to a lesser degree.
The last three chapters deal with AI’s potential future implications and provide recommendations on how societies should proceed. The fifth chapter takes to task the impacts of AI on the international system, relations between states, the conduct of war and diplomacy, and puts forward a few recommendations on how states can mitigate the dangers of AI proliferation. The sixth chapter outlays potential changes to civil and domestic life in the future as AI becomes more prevalent in society. One interesting notion is that attitudes towards AI are likely to change in the future: the current generations who are and will continue to be adult adopters of AI will conceivably differ in their perspectives on the technology than AI natives, in the same way digital natives approach the digital world differently than those who did not grow up with smart phones. The seventh chapter consolidates and builds upon the previous chapters outlining what the authors see as a desperate need for a clearer focus by intellectuals, the citizenry and public servants on AI, to ensure the creation of ethical frameworks and strategic doctrines for AI.
As for the security realm, a central theme of the book is that AI likely will change the incentives policymakers and statesmen face in their dealings with other actors on the international stage. At minimum, AI will alter the weights of inputs in any strategic assessment. One of the important factors that influence incentives is the sheer difficulty in assessing the AI capabilities of other actors and states: it is difficult to assess AI capability in relation to the capabilities of other actors without deploying it, and capabilities may be developed in secret. As a result, states may find it irresistible to act first, rather than to act wisely, as the authors put it.
One aspect that the authors do not address head on are the relevant differences in capabilities between first rate AIs and second-rate AIs. This is a question of technical nature that has strategic implications. Will marginal improvements in military AI technology enable an actor to dominate its rivals on the battlefield? Or will marginal improvements in AI technology be battlefield enablers that can be offset by a relatively small increase in force size? Presumably the answer, either way, will have nontrivial consequences for decisions about funding and strategic doctrine at the highest level.
Yet overall, the authors make a compelling argument that leaders, in most if not all fields, need to think long and hard about the future of society as we enter the age of AI. This is particularly true for the great powers poised for competition in the coming decades, the United States and China. Unsurprisingly, these states are furthest along in their AI development according to the authors, with Russia coming in slightly behind the two giants. Europe, however, is missing in action, which should be cause for concern across the continent, from London to Paris, to Berlin and Rome and ultimately Brussels.
There is still hope for those wanting to catch up. While the technology has perhaps passed infancy, it is still young, and the most impactful developments are probably yet to come. This can be seen in the authors style of recommendations throughout the book, which tend to be of a discursive nature, such as committees to study the issue or to establish mechanisms for international dialogue: that means Rubicon is yet to be crossed and that there is still time to act on AI.
Isak Runarsson is a MPA candidate at Harvard University and Intern Researcher at IAI