Separating hype from reality in the age of ChatGPT is daunting. As experts decry the disruptive and potentially existential challenge of new AI models, headlines regularly trumpet how AI-enabled systems used in Ukraine and the Middle East or developed for U.S.-China competition are indelibly changing warfare. But AI is not yet revolutionizing warfare, and that isn't necessarily a bad thing.

In this tightly written, compelling book, Owen J. Daniels explores the potential and limits of military AI. He argues that, rather than constituting entirely new ways of war, advancements in artificial intelligence are helping militaries to finally realize long-foreseen uses for the technology, uses that even Cold War technologists incorrectly predicted would reshape warfare in their own era. The distinction between a military AI which is revolutionary or evolutionary is far from pedantic, Daniels argues. At present, long-standing limitations – from the complexity of human-machine teaming to the difficulty of navigating dynamic battlefields – hinder any fundamental change to the character of warfare. Global leaders have a narrowing window to shape its trajectory by developing safe, governable AI systems and international understanding around their responsible use. Otherwise, AI could become the threat we all fear.

Owen J. Daniels is Associate Director of Analysis and an Andrew W. Marshall Fellow at Georgetown's Center for Security and Emerging Technology (CSET).