Artificial intelligence poses a growing threat that may, in some respects, surpass the strategic risks associated with nuclear weapons, according to discussions at a high-level defence forum in Singapore, where military leaders and humanitarian officials warned that rapidly accelerating technologies could compress decision-making cycles and increase the likelihood of miscalculation in conflict situations.
At the Shangri-La Dialogue, participants said AI-driven systems are shortening what military strategists describe as the “OODA loop” — observe, orient, decide and act — to the point where human judgment may no longer be able to keep pace with battlefield developments. Lieutenant General Nauman Zakria, Commander of 1 Corps and Army Rocket Force Command of the Pakistan Army, warned that such compression creates a form of operational fog in which leaders may struggle to properly assess unfolding situations.
“When you collapse the OODA loop, a human can’t evaluate the situation fast enough,” Zakria said during a panel discussion on Saturday. He cautioned that under such conditions, decision-making could become impulsive and unstable, increasing the risk of escalation. “People will act irrationally, and the actions will be extreme,” he added, highlighting concerns that artificial intelligence may reduce the time available for measured military responses.
Those concerns were echoed by Mirjana Spoljaric, president of the International Committee of the Red Cross, the only panelist not directly representing a military institution. She emphasised that while technological innovation can deliver humanitarian benefits, it also introduces unprecedented risks in modern warfare, particularly when it is integrated into weapons systems and command structures.
“We don’t know where the trigger is pulled,” Spoljaric said. “It could be thousands of kilometres away. So while there are potentials of AI for protecting civilians, what we see at the moment is only the negative side.” Her remarks underscored growing unease among humanitarian organisations about the use of autonomous or semi-autonomous systems in conflict zones, where accountability and human oversight may be diminished.
Despite the focus on artificial intelligence, nuclear weapons remained a central reference point in the debate over strategic stability. Major General Meng Xiangqing of the People’s Liberation Army reiterated China’s longstanding no-first-use nuclear policy, arguing that broader adoption of similar doctrines could contribute to reducing global tensions and strengthening deterrence frameworks.
“If we can do so, we can reduce the risk and we can further enhance strategic stability,” Meng said, framing nuclear restraint as a continuing pillar of international security architecture even as new technologies emerge.
However, much of the discussion repeatedly returned to artificial intelligence and its accelerating role in modern warfare. General Onno Eichelsheim, the Netherlands’ chief of defence, noted that AI is no longer theoretical in military planning but is already actively deployed on contemporary battlefields, including in Ukraine, where it has been used to help anticipate Russian military movements. The United States has also acknowledged incorporating AI tools into operational planning for strikes on Iranian targets, further illustrating its expanding military application.
“AI is a huge risk in escalation. I think that’s clear,” Eichelsheim said. “But I’m not naïve. It’ll be used in the domain. It is already being used.” His comments reflected a broader consensus among defence officials that, despite its risks, AI integration in military systems is inevitable and already underway across multiple theatres of conflict.
The debate in Singapore highlighted a growing tension between technological advancement and strategic stability, as governments and defence institutions grapple with how to regulate and control systems that can process information and generate decisions far faster than human operators. Participants warned that the acceleration of warfare decision cycles could increase the probability of unintended escalation, particularly in high-pressure scenarios where states must respond rapidly to perceived threats.
Analysts at the forum suggested that while nuclear deterrence remains a cornerstone of global security doctrine, artificial intelligence introduces a different category of risk — one defined not by destruction alone, but by speed, ambiguity and the erosion of human control in critical decision-making chains.
As discussions concluded, participants agreed that future security frameworks will need to account for both traditional deterrence models and emerging technological systems. However, there was little consensus on how such regulation might be achieved, reflecting the complexity of integrating AI governance into existing international security structures.
The Shangri-La Dialogue, one of the world’s most prominent defence summits, has increasingly become a venue for addressing the intersection of technology and warfare. This year’s discussions underscored that while nuclear weapons remain a defining feature of strategic balance, artificial intelligence is rapidly emerging as a parallel concern — one that may reshape the nature of conflict in the decades ahead.
