Iran war’s quiet warning: AI is rewriting the nuclear rulebook

The ongoing US-Israel-Iran conflict symbolises the paradigmatic shift in the conduct of warfare. The use of artificial intelligence (AI) on the battlefield, first seen experimentally during the Russia-Ukraine War, has now matured to form an integral element of US and Israeli kinetic operations targeting Iran. Understanding this transition to an AI-driven battlefield is essential for all militaries, particularly for nuclear-armed states like India, which are recalibrating their strategic postures amid accelerating technological change.

In the Russia-Ukraine conflict, both sides tested AI-driven capabilities such as long-range targeting and multi-domain operations against each other. The US and Israel are now applying the lessons drawn from that conflict to a more critical and consequential theatre of Iran. The deployment of AI tools to expedite decision-making by automating elements of drone systems, satellite imagery analysis, target identification and battle simulations demonstrates this shift. Platforms like Project Maven, the Microsoft-led Integrated Visual Augmentation System, Claude LLM, and Amazon Web Services have become integral to US military planning and campaign execution.

The most consequential change is the compression of the sensor-to-shooter loop, reducing the time between intelligence collection and strike execution, with processing of over a thousand targets from days to minutes. This marks a profound change in the tempo and scale of warfare. However, it also raises a fundamental question about the scope of human judgment and accountability.

The implications of this shift reverberate beyond the immediate battlefield. AI’s role in military decision-making could reshape the dynamics of both vertical and horizontal escalation among warring parties in complex ways that are difficult to control or even predict. It is pertinent to remember that all AI systems carry a statistical probability of error (besides ‘hallucination’), and hence even a single miscalculation in targeting, such as what was seen in the strike on a girls’ school in Minab, Iran, carries grave implications – not just due to collateral damage but also escalation dynamics. In a potentially worst-case scenario, greater autonomy and reduced human oversight could lead the AI system to recommend high-value targets whose strike might trigger unintended responses, blurring the distinction between conventional and nuclear thresholds.

Quo Vadis India?

This has implications for nuclear deterrence. The risk is that not only is AI making warfare more destructive, but it is also complicating crisis management through inadvertent escalation, potential miscommunication, and therefore signalling to the adversary.

For these reasons, India cannot be a mute observer of these shifts in warfare. It is the only nuclear-armed state engaged in active military disputes with two nuclear-armed neighbours that share a close defence and security bond. Both China and Pakistan are actively integrating AI into their military and nuclear architectures. For instance, China is developing AI-enabled targeting and early-warning capabilities within its NC3 architecture as part of its nuclear force modernisation, to strengthen deterrence vis-à-vis the United States.

Given these developments, the Indian military must conduct an integrated assessment of its nuclear posture to address its adversaries’ rapid integration of AI capabilities. New Delhi’s normative position on this issue, anchored on human judgment and oversight in the military application of AI, is fine for global norms-building. However, taking realpolitik into consideration, it must study the impact of AI-driven adversarial capabilities on its second-strike deterrent. The question of how AI-driven adversarial momentum affects India’s nuclear signalling, response timelines, and escalation calculus needs an immediate and greater scrutiny. This line of inquiry should become a priority for today to serve the realities of tomorrow.

(Sameer Patil is Director, Centre for Security, Strategy and Technology, Observer Research Foundation (ORF), Rahul Rawat is a Research Assistant in ORF’s Strategic Studies Programme)