- India emphasised that human accountability must remain central in military AI decisions at a UN session
- Ambassador Charanjeet Singh stressed the need to preserve human oversight throughout AI's military lifecycle
- Singh urged equitable AI access for developing countries and cautioned against technological divides
India told a United Nations gathering on military applications of artificial intelligence that accountability for decisions on the use of force must stay with humans, warning that such responsibility "cannot be transferred to machines".
Speaking at informal exchanges on artificial intelligence in the military domain held in Geneva on June 15, Indian Ambassador Charanjeet Singh said human judgement and oversight must be preserved "throughout the AI lifecycle", even as armed forces worldwide rapidly fold AI into their operations. The session was convened by the UN Office for Disarmament Affairs along with the Netherlands and South Korea, with India serving as a co-sponsor.
The talks come against the backdrop of growing international concern that military AI adoption is outpacing the rules meant to govern it. In December 2024, the UN General Assembly passed its first-ever resolution on the subject, 79/239, which on the recommendation of the First Committee affirmed that international law applies across every stage of a military AI system's life cycle and called on states to confront the opportunities and risks involved. The resolution also directed the Secretary-General to collect member states' views on the issue.
India was among 30 countries that contributed to the resulting Secretary-General's report. Singh said those submissions reflected "universal agreement" that international humanitarian law – particularly the principles of distinction, proportionality and precaution – applies to military AI and that nearly every country flagged the danger of advanced AI capabilities falling into the hands of non-state actors, risking "misunderstandings, miscalculations and misperceptions". Despite that consensus on the need for governance, he acknowledged states remain split over which forum should set the rules, even as the report offers a possible starting point for talks.
Much of Singh's address focused on equity for the developing world. He cautioned that emerging AI governance frameworks must not entrench existing technological divides, and that any future rules should ensure "equitable access" to AI so its benefits reach all countries, "particularly developing countries", while avoiding the stigmatisation of the technology itself. Concerns over unequal access and participation in governance discussions, he said, "warrant careful consideration."
On the more contested question of autonomous weapons, Singh said discussions must stay grounded in "operational realities" and respect national jurisdictions while still allowing risks to be identified and managed. He pointed to a framework India has developed for "trustworthy AI in defence", built around reliability, safety, transparency, fairness and privacy, which he said could inform wider international deliberations; New Delhi hosted a side event in Geneva the same day to present it. He also noted India's continuing role in the UN's Group of Governmental Experts on Lethal Autonomous Weapons Systems under the Convention on Certain Conventional Weapons, including its contribution to the group's 11 Guiding Principles.
Singh warned, however, that the growing number of parallel tracks – the LAWS process alongside an expanding set of plurilateral initiatives – carries a "real risk of fragmentation" and competing rulebooks that could dilute international engagement. Such informal and plurilateral efforts, he argued, should "reinforce and enrich" established multilateral processes rather than substitute for them.
He closed by saying he hoped the Geneva exchanges would "help build greater convergence" among delegations navigating one of the most consequential – and least regulated – frontiers in global security policy.
