By Stuart Kerr, Technology Correspondent, LiveAIWire
AI military strategy is not a projected capability, it is an operational reality already reshaping how major powers plan and fight wars. Military strategy has always been shaped by the technologies available to the forces that pursue it. The introduction of gunpowder, the internal combustion engine, and nuclear weapons each fundamentally altered the logic of conflict. Artificial intelligence is the latest in this sequence, and there is broad agreement among defence analysts that its implications are at least as significant as any of its predecessors.
The integration of AI into military systems is already well advanced: autonomous weapons, AI-powered intelligence analysis, machine learning logistics, and algorithmic cyber operations are operational realities in several major military powers, not projected capabilities.
Intelligence and Surveillance: The Data Advantage
The most consequential near-term military application of AI military strategy is in intelligence, surveillance, and reconnaissance. Modern conflicts generate enormous volumes of sensor data: satellite imagery, signals intercepts, drone feeds, and open-source intelligence from social media and communications networks. Human analysts cannot process this information at the speed required for operational decision-making. AI can.
Project Maven, the US Department of Defense programme launched in 2017 to apply computer vision to drone footage analysis, demonstrated that machine learning could identify objects of military interest at a rate and accuracy that human analysts could not match. This same surveillance capability, deployed at civilian scale, is the subject of LiveAIWire’s coverage of how AI drone surveillance is piloting the future of the sky, where border agencies and police forces already use the technology this military programme helped establish.
Autonomous Weapons: The Lethal Decision Problem
The most ethically fraught dimension of AI military strategy is autonomous weapons: systems capable of selecting and engaging targets without direct human control of the lethal decision. The International Committee of the Red Cross has called for binding international regulation of autonomous weapons systems, arguing that the decision to use lethal force must remain with a human who can assess proportionality, distinguish combatants from civilians, and bear legal responsibility for the decision.
Critics argue further that the accountability vacuum created by removing human decision-making from the lethal act creates unacceptable impunity: if a machine kills a civilian, who is responsible? The programming decisions that led to the outcome may be buried in millions of lines of code, made years before the incident by engineers who did not anticipate the specific scenario. This is a version of the same accountability gap LiveAIWire has traced in our coverage of AI sentencing bias in predictive risk tools, where consequential decisions about people are delegated to systems whose reasoning cannot be fully audited.
Cyber Operations: The Invisible Battlefield
AI is transforming cyber warfare in ways that extend far beyond traditional hacking. Machine learning systems can identify vulnerabilities in target networks at speeds human operators cannot, generate and adapt malware in response to defensive countermeasures, and conduct influence operations at a scale that earlier disinformation campaigns could not achieve. The NATO Cooperative Cyber Defence Centre of Excellence has documented the integration of AI into offensive and defensive cyber operations across allied and adversary militaries. Attribution of AI-assisted cyber attacks is more difficult than attribution of traditional attacks, reducing the deterrent effect of response threats.
Governance: Racing Ahead of the Rules
International governance of AI military strategy is at an early stage. Discussions within the Convention on Certain Conventional Weapons have been ongoing since 2014 without producing binding regulation. The United States Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy has been endorsed by over 50 countries, but is non-binding and contains no verification mechanism. Russia and China have not endorsed it. The consequence is that the rules governing AI military strategy are being determined primarily by the deployment decisions of major military powers rather than by international law, echoing the same governance lag LiveAIWire has traced in our coverage of AI diplomacy and how nations are using code as a new form of soft power.
The Human Dimension: Soldiers, Commanders, and Machines
Military AI raises questions not only about international law and strategic advantage but about the human experience of war. Research on human-machine teaming in military contexts has identified consistent patterns: operators tend to over-trust automated systems, particularly when those systems have previously been accurate, leading to acceptance of system recommendations without adequate critical assessment.
When a commander accepts an AI-generated targeting recommendation that results in civilian casualties, the legal and moral responsibility for that outcome does not transfer to the machine. Understanding the limitations and error modes of AI systems is now a core competency for military commanders, in the same way that understanding the capabilities and limitations of a specific weapons system has always been.
AI Military Strategy in Simulation and Wargaming
One of the less-discussed but practically significant applications of AI military strategy is in simulation and wargaming. AI systems can generate and evaluate millions of simulated conflict scenarios far faster than human wargaming teams, identifying strategic vulnerabilities and testing the robustness of operational plans against a range of adversary responses. The US and UK militaries have invested substantially in AI-enhanced simulation environments.
The risk of over-reliance on simulation outputs is analogous to the automation bias problem in operational contexts. A system that consistently recommends a particular operational approach because it succeeded in a large number of simulated scenarios may be training planners to favour an approach that is well-optimised for the simulated world but vulnerable in the actual one. Maintaining genuine critical engagement with simulation outputs, rather than treating them as authoritative strategic guidance, is a professional responsibility that AI military strategy adoption makes more rather than less important.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity, and the social impact of emerging technology. He publishes daily at LiveAIWire.com.