Security in age of AI warfare requires capability and governance

Audio report: written by reporters, read by AI

Yoon Byung-se

The author is a former foreign minister of Korea and former chair of the REAIM Global Commission on Responsible AI in the Military Domain.

“Humanity must not entrust its fate to the black box of algorithms.”

Despite this warning by United Nations Secretary General António Guterres, the international order is already shifting toward a reality in which artificial intelligence is becoming a decisive factor in warfare. Conflicts such as the Russia-Ukraine war, the Israel-Hamas war and the recent U.S.-Israel confrontation with Iran demonstrate how AI is being integrated into key stages of the kill chain. Compared to the precision warfare of the Gulf War or the network-centric operations of the Iraq War, AI now plays a coordinating role on the battlefield, reshaping both the speed and structure of war.

In the Iran conflict, AI fuses data from satellites, signals intelligence, drones and radar in real time, compressing the kill chain to minutes. This enables simultaneous implementation of multidomain operations under frameworks such as Combined Joint All-Domain Command and Control. The ability to strike thousands of targets in a short period reflects this structural transformation. It also aligns with the evolution of the Maven project initiated during the first U.S. President Donald Trump administration, expanded under former U.S. President Joe Biden and further accelerated in its battlefield application during Trump’s second term. By combining “data integration” with “meaning integration,” AI is redefining how wars are conducted.

After the Sept. 11 attacks, a bipartisan commission criticized intelligence agencies for failing to “connect the dots.” Today, AI is increasingly addressing that gap. Information gathered through multiple channels is integrated into a common operational picture, while the kill chain is evolving into a “kill web” that incorporates cyberoperations to disable adversaries. At the Responsible AI in the Military domain (REAIM) summit hosted by Korea in September 2024, the United States outlined six domains for military AI application: intelligence, surveillance and reconnaissance, command and control, logistics, human-machine teaming, simulation and operations in space, cyber and electromagnetic. These domains underscore AI’s central role in modern warfare.

The U.S. Department of Defense now refers to “algorithmic warfare” and “AI-enabled battlefield operations,” while China promotes its doctrine of what it calls “intelligentized warfare.” Discussions in Washington increasingly center on the arrival of an AI war era.

Examples from recent conflicts further illustrate this shift. Drone swarms and loitering munitions in Ukraine, as well as systems such as Lavender and Gospel used in Gaza, show how AI can dramatically accelerate targeting and operational tempo. Even at an early stage, military AI is already shaping the pace of war beyond conventional weapons systems, while the scope of human decision-making continues to narrow. The temptation to move toward fully autonomous weapons is likely to grow.

These changes are elevating AI competition to the core of military rivalry. Victory in unseen domains increasingly determines outcomes on the battlefield. Some analysts argue that AI capabilities now carry strategic significance comparable to tactical nuclear weapons. Former U.S. Secretary of State Henry Kissinger once predicted that AI would shape the global order of the 21st century, a forecast that is now approaching reality.

Yet the pace of military AI development far outstrips global governance efforts. Intergovernmental discussions on lethal autonomous weapons systems, United Nations (UN) resolutions and REAIM initiatives continue, but the gap between norms and reality has widened during Trump’s second term. A central issue is how to maintain human control and define the acceptable scope of autonomy. Recent tensions involving the U.S. Department of Defense and the AI firm Anthropic, which was designated as a supply chain risk entity, reflect differing perspectives on these issues.

As AI-driven decision-making accelerates to machine speed, strategies such as “left of launch,” aimed at neutralizing threats before launch, are becoming more feasible. At the same time, the risk of miscalculation and civilian casualties is increasing. A study led by Prof. Kenneth Payne for the REAIM Global Commission found that AI models chose nuclear use in 20 out of 21 simulated conflicts. This highlights the risks inherent in delegating critical decisions to machines. The commission’s final report emphasized maintaining human control over decisions involving nuclear weapons.

At a summit in late 2024, then-President Biden and Chinese President Xi Jinping agreed that humans must retain ultimate control over decisions related to nuclear weapons. This principle should be reaffirmed in future bilateral meetings.

On the Korean Peninsula, the implications of AI warfare are particularly significant. Shorter response times require a fundamental redesign of crisis management systems. As Korea pursues the transfer of wartime operational control, it must secure not only next-generation weapons but also capabilities in C4ISR, or AI-based Command, Control, Communications, Computers, Intelligence, Surveillance and Reconnaissance, as well as in cyberoperations and integrated manned-unmanned systems. These elements define the “invisible war.”

President Lee Jae Myung emphasized this balance while presiding over a UN Security Council meeting in September of last year. He noted that AI could become either a dangerous force or a beneficial tool, urging the international community to establish principles for responsible use.

Korea, which aspires to become a global leader in AI, must strengthen its military capabilities while embedding responsibility and safety from the design stage. The National Intelligence Service’s recent initiative with seven countries on “security by design” for AI supply chains reflects this direction. It aligns with Korea’s broader diplomatic and security strategy, which seeks to shape a digital international order through platforms such as the UN and REAIM.

Ultimately, Korea must pursue a comprehensive strategy that integrates government, military and private-sector expertise. Strengthening national security and prosperity requires not only technological advancement but also a commitment to ensuring that AI serves humanity.

This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.