The character of contemporary warfare is undergoing a profound shift. The intersection of artificial intelligence, distributed data science, offensive cyber capabilities and autonomous systems has evolved beyond mere technological iteration and has formed a volatile strategic nexus that challenges the foundational architecture of international law.
We are witnessing a dangerous divergence between the rapid acceleration of military capabilities and the slow evolution of legal norms. Most states are moving to rely increasingly on decentralised information networks and automated decision-making frameworks. The ultimate challenge of modern defence is no longer merely a race for advancement, but ensuring that lethal weapons operate under the principles of state responsibility and human accountability.
Distinction, Proportionality and Military Necessity
It is important to return to the core tenets enshrined in the Geneva Conventions and their Additional Protocols. The integration of advanced algorithms into target selection and weapon deployment directly threatens the bedrock principles of distinction, proportionality and military necessity.
International humanitarian law dictates that combatants must at all times distinguish between civilian populations and military objectives, and that any anticipated civilian harm must not be excessive in relation to the concrete and direct military advantage anticipated. Autonomous systems reliant on machine learning models trained on historical datasets lack the capacity for situational nuance, moral reasoning and contextual empathy required to make these delicate legal evaluations on a dynamic battlefield. Legal judgment is converted into a game of probabilities, which strips the law of its humanitarian core.
The Accountability Vacuum
The deployment of autonomous systems also introduces an unprecedented accountability vacuum. Liability used to trace back to a human actor: a commanding officer who authorised an attack, or a political leader who sanctioned a campaign. An algorithmic system can err, misinterpret sensor data or exhibit emergent behaviour in a complex environment, which obscures the line of causation.
The question remains: can a military commander be held liable for a war crime committed by an unexplainable network of large datasets? It has been argued that if the system’s decision-making process is opaque, then proving the requisite mental element for international crimes becomes impossible. Reliance on automated systems therefore risks creating a vacuum in which violations of international law occur with impunity, undermining the deterrent effect of international jurisprudence.
Intelligence Agencies as Norm Entrepreneurs
This legal ambiguity is compounded by the evolving behaviour of state intelligence agencies, which increasingly act as de facto norm entrepreneurs in cyberspace. International law has always been forged through formal mechanisms: the drafting of multilateral treaties, the codification of state practice and, over time, the slow crystallisation of opinio juris.
Informal institutional practices are now outpacing formal diplomacy. As intelligence agencies conduct state-sanctioned cyber operations under the cloak of deniability, their operational conduct establishes practical baselines for acceptable state behaviour. If a specific class of cyber attack or digital espionage goes unpunished or becomes widely practised, it shapes international expectations and customs long before diplomats ever convene. This bottom-up method of lawmaking risks normalising destabilising behaviours such as targeting critical civilian infrastructure or compromising democratic institutions, thereby weakening the traditional framework of state sovereignty.
Reasserting Control: Cyber Diplomacy and Structural Legal Innovation
There is a need to overcome the fatalistic belief that technology dictates strategic outcomes and that law must simply adapt ex post facto. The international community must reassert control through proactive cyber diplomacy and structural legal innovation. Cyber diplomacy should not be viewed as an idealistic alternative to strategic competition, but as an essential and complementary mechanism to manage geopolitical risk.
The lessons of the nuclear age are instructive. That era necessitated bilateral arms control treaties and hotlines to prevent catastrophic miscalculation. The cyber and AI era demands robust frameworks for information sharing, crisis communication and consensus-based limits on algorithmic warfare. Diplomacy must be used to codify explicit red lines, such as a categorical prohibition on removing human judgment from the loop of nuclear command and control and from lethal autonomous targeting.
Article 36 and the Case for Explainability
Bridging the gap between technical execution and legal compliance requires interdisciplinary work. Military organisations must integrate legal advisers directly into the development, training and deployment phases of data science tools and AI architectures, with focused attention on the operational risks of algorithmic bias, data corruption and system failure.
Compliance with accountability principles requires the development of explainable AI systems that allow legal reviewers to reconstruct the precise rationale behind automated recommendations. If a technology cannot be subjected to rigorous legal review under Article 36 of Additional Protocol I to the Geneva Conventions, which has long mandated that states determine whether the employment of a new weapon has the potential to violate international law, then it should not be permitted on the battlefield.
Conclusion
The integration of AI, cyber operations and data science does not eliminate the inherent uncertainty of warfare. It merely shifts that uncertainty from the tactical sphere to the domains of governance, ethics and international regulation. The measure of a state’s power in the twenty-first century will not be determined by the sophistication of its code or the speed of its processors, but by the resilience of its political, legal and institutional frameworks. It is only by reinforcing the rule of law that the digital transformation of warfare can be prevented from devolving into an era of unchecked devastation.
Esquare Legal advises technology companies and institutions on AI governance, regulatory exposure and cross-border compliance. See our Think Tank, or contact us to discuss a mandate.
Author: Hassan Raza, Senior Associate, Esquare Legal.