Once limited to research labs, artificial intelligence is now deeply ingrained in military planning, surveillance, logistics, and the development of weaponry. As governments increasingly incorporate AI into national security strategies, the main concern is not whether AI will influence warfare, it already does but whether international law can change swiftly enough to control its application.
Discussions at the UN, talks under the Convention on Certain Conventional Weapons (CCW), and the Responsible AI in the Military Domain (REAIM) initiative have all contributed to the urgency of this subject. Together, these approaches indicate growing realisation that military AI requires substantial legal regulation before technical improvements outrun international standards.
There is no denying the military benefits of AI. It can handle enormous volumes of data, enhance situational awareness, bolster intelligence analysis, and facilitate quick decision-making. The biggest worry is when AI starts to replace human judgement instead of just assisting it. The foundation of international humanitarian law is human judgement.
Algorithms are unable to replicate the context, legal reasoning, and accountability required by the principles of distinction, proportionality, military necessity, and precaution. AI is incredibly quick at processing data and recognising patterns, but it is not able to make moral decisions or be held legally accountable for wrongdoing.
As a result, real human control must become a legal obligation rather than a political ideal. Human monitoring should be crucial to every choice involving lethal force. When autonomous systems are allowed to identify and engage targets independently, responsibility becomes increasingly ambiguous.
Under international law, command responsibility cannot be replaced by software faults, biased datasets, cyber manipulation, or algorithmic flaws. When liability is ambiguous, compliance with armed conflict laws weakens.
Regulation should, however, refrain from impeding acceptable technical advancement. Healthcare, agriculture, scientific research, disaster management, and climate resilience are all aided by artificial intelligence.
Governance should prioritise preventing misuse while guaranteeing fair access for peaceful objectives because these technologies are by nature dual-use. Developing nations should not be shut out of technology growth through discriminatory governance practices or excessively stringent export regulations.
This issue is especially pertinent to the Global South. Technologically sophisticated states that already have advanced AI capabilities continue to shape a large portion of the present debate. Their choice for voluntary rules may minimise the influence of nations that are still building their own technology capabilities while maintaining flexibility for themselves.
Therefore, all governments should be able to contribute significantly to the development of the legal regulations controlling military AI under a lawful international framework.
Pakistan has long supported such an approach, calling for a legally enforceable instrument that prohibits fully autonomous weapons systems from operating without significant human control, as well as promoting the peaceful use of new technology and equal access to scientific innovation. This approach underlines the necessity to strike a balance between humanitarian assistance and technological advancement.
Despite growing international attention, existing governance remains fragmented. The CCW provides the principal negotiating forum but has struggled to achieve consensus.
UN General Assembly resolutions have generated political momentum without creating binding legal obligations, while the REAIM process has broadened dialogue among governments, industry and academia without establishing enforceable standards. The result is a patchwork of voluntary commitments that falls short of a comprehensive legal framework.
A practical approach should combine prohibition with regulation. Fully autonomous weapons capable of selecting and attacking human targets without meaningful human control should be prohibited.
Other AI-enabled military systems should remain subject to rigorous legal review, cybersecurity requirements, transparency standards, testing, explainability and clear chains of command responsibility. Such safeguards would preserve innovation while ensuring compliance with international humanitarian law.
The strategic implications become even more significant when AI intersects with nuclear command, control and communications (NC3). AI-assisted early warning, intelligence fusion and decision-support systems may improve efficiency, but they also compress decision-making timelines and increase the risks associated with false alarms, corrupted data and cyberattacks.
In nuclear deterrence, speed is not always an advantage. Decisions involving strategic forces require verification, political judgment and human accountability, qualities that algorithms cannot provide.
For South Asia, where technological competition already exists alongside nuclear deterrence, these risks deserve particular attention. AI-enabled surveillance, cyber operations and autonomous systems could increase uncertainty during crises and shorten opportunities for political decision-making.
Regional confidence-building measures, including a code of conduct reaffirming meaningful human control, improving crisis communication and promoting dialogue on AI safety and cybersecurity, could help reduce the risks of unintended escalation.
International law has often developed only after the consequences of new weapons became evident. Military AI presents an opportunity to act before that pattern repeats itself. The future of warfare should not be determined solely by technological capability but by legal responsibility and human judgment.
Preserving meaningful human control, strengthening accountability and safeguarding strategic stability are essential if artificial intelligence is to remain a tool that serves humanity rather than one that weakens the legal principles governing the use of force.
