The Speed of Innovation vs. The Pace of Diplomacy
Military technology is advancing at an astonishing rate. By 2026, AI is no longer a futuristic concept but a standard component in defence systems, from AI-augmented decision-making on the battlefield to autonomous mission planning. AI is being used to fuse
data from surveillance, logistics, and command-and-control platforms to enable faster, more intelligent actions. This rapid operational deployment, seen in conflicts in Ukraine and Gaza, is fundamentally changing the character of warfare. In contrast, international governance, which relies on consensus-building among nations, moves at a deliberative, often glacial, pace. Forums like the UN's Convention on Certain Conventional Weapons (CCW) have been discussing lethal autonomous weapons systems (LAWS) since 2014, but progress toward a legally binding treaty has been slow and fraught with disagreement. This mismatch in speed is the central reason for the growing regulatory gap.
The National Security Impasse
When it comes to defence, nations are notoriously protective of their sovereignty and strategic advantages. Major military powers fear that agreeing to binding restrictions on AI weapons will leave them vulnerable to rivals in what has become a new AI arms race. This strategic competition discourages restraint. The secrecy inherent in advanced military research and development creates a further obstacle. Without transparency, it is nearly impossible to verify compliance with any proposed treaty. Recent talks have seen key nations like the US and Russia weaken proposed language that would mandate human review of AI-selected targets, prioritising the rapid deployment of these technologies over multilateral consensus. This highlights how national security interests consistently override calls for collective governance.
The Dual-Use Dilemma
A significant hurdle in regulating military AI is the 'dual-use' problem. Many AI technologies, such as advanced semiconductors, quantum computing, and machine learning algorithms, have both civilian and military applications. An algorithm developed for medical diagnostics could be repurposed for target recognition, or a drone designed for logistics could be adapted for surveillance or attack. This makes it incredibly difficult to craft regulations that ban or restrict 'weaponised AI' without also stifling innovation in the commercial sector. Unlike nuclear materials or chemical agents, the building blocks of AI are often software-based and can be moved around the world with little trace, making any regulatory regime incredibly difficult to enforce. This reality complicates any attempt to draw clear lines around what constitutes a prohibited system.
A Fractured Global Consensus
There is no universal agreement on how to govern military AI, or even a common definition of what constitutes a 'lethal autonomous weapon'. While an overwhelming majority of 164 nations at the UN have voted to move forward on addressing the issue, key military powers remain hesitant. Concerns that machines are on track to 'crowd out' human decision-making in warfare have been repeatedly raised by the UN Secretary-General and disarmament experts. India, for its part, has taken a firm stance that the decision to use nuclear weapons must remain firmly in human hands, acknowledging the risks of AI integration in command-and-control systems. New Delhi has also called for a global non-proliferation initiative for AI weapons systems. However, with different blocs of nations holding fundamentally different views on ethics, reliability, and the necessity of 'meaningful human control', achieving a global consensus remains a distant goal.
