Engines of Power: Electricity, AI, and General-Purpose Military Transformations

Jeffrey Ding, Allan Dafoe

Abstract

Major theories of military innovation focus on relatively narrow technological developments, such as nuclear weapons or aircraft carriers. Arguably the most profound military implications of technological change, however, come from more fundamental advances arising from general purpose technologies (GPTs), such as the steam engine, electricity, and the computer. Drawing from the economics literature on GPTs, this paper distills several propositions on how and when GPTs affect military affairs, a phenomenon termed General-Purpose Military Transformations (GMTs). We argue that the impacts of GMTs on military effectiveness are broad, delayed, and shaped by indirect productivity spillovers. Additionally, GMTs differentially advantage those militaries that can draw from a robust industrial base in the GPT. Through a case study of electricity and an application to artificial intelligence, we illustrate the profound implications of these transformations.

Redefining Military Innovation

Most political science theories regarding military innovation center on "narrow" technologies—specific inventions like the tank or the submarine. While these are undoubtedly impactful, they do not capture the systemic shifts caused by General-Purpose Technologies (GPTs). A GPT is defined by three core characteristics: pervasiveness across sectors, a high potential for continual improvement, and "technological complementarities"—the ability to make other technologies more effective.

When a GPT enters the military sphere, it triggers a General-Purpose Military Transformation (GMT). Unlike narrow innovations, GMTs do not follow a standard arms-race model. Instead, they reshape the entire foundation of military power, often in ways that are not immediately visible on the battlefield.

The Mechanics of GMTs: Broad, Delayed, and Indirect

Our theory proposes that GMTs differ from narrow technological shifts in three critical ways:

  1. The Gestation Paradox: GPTs require decades to realize their full potential. In the civilian sector, electricity did not significantly boost factory productivity until the 1920s—nearly 40 years after its introduction—because it required a total redesign of factory floors from single-engine steam power to distributed electric motors. Military GMTs face similar delays, as they require deep organizational restructuring and the "upskilling" of personnel.
  2. Indirect Productivity Spillovers: The most significant military advantages of a GPT often occur far from the front lines. For example, the electrification of industrial manufacturing allowed for the mass production of aircraft and munitions during the World Wars, shifting the global balance of power through industrial capacity rather than just tactical superiority.
  3. Systemic Integration: Over time, a GPT permeates every aspect of military hardware. By 1960, for instance, electrical installations accounted for 30% of the total cost of British warships, illustrating how a GPT eventually becomes the backbone of all military platforms.

The Industrial Dependency Hypothesis

A central finding of our research is the "Industrial Dependency Hypothesis": a military’s ability to successfully exploit a GPT depends heavily on its connection to a robust domestic civilian industrial base.

To test this, we compared the adoption of electricity (a GPT) with the submarine (a narrow technology) in the early 20th century. Russia provides a cautionary tale: despite pioneering electronic countermeasures as early as 1904, Russia failed to electrify its military effectively because it lacked a domestic civilian electrical industry. In contrast, Great Britain and the United States succeeded by tapping into a vast pool of civilian talent, infrastructure, and innovation. Interestingly, Russia remained competitive in submarine technology during the same period, suggesting that while "laggard" nations can keep pace with narrow technologies, they struggle with the broad requirements of a GMT.

AI as the "New Electricity"

Applying these historical lessons to the modern era, we argue that Artificial Intelligence is the contemporary equivalent of electricity. Current discourse often focuses on "narrow" AI applications, such as autonomous drones or "killer robots." However, if AI is truly a GPT, its most profound impact will be a process of "intelligentization"—a slow, broad transformation affecting logistics, human resources, and strategic decision-making.

If AI follows the historical trajectory of electricity, we should expect a significant lag in its military impact. Given that the current deep learning boom began in the 2010s, the full realization of an AI-driven General-Purpose Military Transformation may not occur until the 2050s. Success in this transformation will likely depend less on specific military prototypes and more on a nation's ability to integrate civilian AI advancements into its broader defense infrastructure.

Conclusion

Understanding the distinction between narrow innovations and General-Purpose Military Transformations is vital for defense policy. GMTs are not simply bought; they are cultivated through long-term investment in industrial capacity and organizational adaptation. As we move further into the age of AI, the nations that maintain a lead will be those that recognize AI not just as a new weapon, but as a fundamental engine of power that requires a whole-of-society industrial foundation.

Summary updated

Abstract

Major theories of military innovation focus on relatively narrow technological developments, such as nuclear weapons or aircraft carriers. Arguably the most profound military implications of technological change, however, come from more fundamental advances arising from general purpose technologies, such as the steam engine, electricity, and the computer. With few exceptions, political scientists have not theorized about GPTs. Drawing from the economics literature on GPTs, we distill several propositions on how and when GPTs affect military affairs. We call these effects general-purpose military transformations. In particular, we argue that the impacts of GMTs on military effectiveness are broad, delayed, and shaped by indirect productivity spillovers. Additionally, GMTs differentially advantage those militaries that can draw from a robust industrial base in the GPT. To illustrate the explanatory value of our theory, we conduct a case study of the military consequences of electricity, the prototypical GPT. Finally, we apply our findings to artificial intelligence, which will plausibly cause a profound general-purpose military transformation.
BibTeX
@inproceedings{ding2023engines,
  author = {Jeffrey Ding and Allan Dafoe},
  title = {{Engines of Power: Electricity, AI, and General-Purpose Military Transformations}},
  booktitle = {European},
  year = {2023},
  doi = {10.1017/eis.2023.1},
}
Figure