Episode Summary

In this episode of Interpreting India, host Charukeshi Bhatt is joined by Shimona Mohan, associate researcher at the United Nations Institute for Disarmament Research (UNIDIR). Together, they unpack the rapidly evolving landscape of artificial intelligence in the military domain, with a special focus on lethal autonomous weapons systems (LAWS).

The discussion traces how AI’s dual-use nature complicates governance, highlights the risks of bias and miscalculation, and explores why progress in international negotiations has been slow despite nearly a decade of debate.

Episode Notes

The episode opens with Bhatt framing the global stakes: from drones on the battlefield to AI-powered early warning systems, militaries worldwide are racing to integrate AI, often citing strategic necessity in volatile security environments. Mohan underscores that AI in conflict cannot be characterized in a single way, applications range from decision-support systems and logistics to disinformation campaigns and border security.

The conversation explores two categories of AI-related risks:

• Inherent risks: design flaws, bias in datasets, adversarial attacks, and human–machine trust calibration.
• Applied risks: escalation through miscalculation, misuse in targeting, and AI’s role as a force multiplier for nuclear and cyber threats.

On governance, Mohan explains the fragmentation of current disarmament processes, where AI intersects with multiple regimes, nuclear, cyber, conventional arms, yet lacks a unified framework. She highlights ongoing debates at the UN’s Group of Governmental Experts (GGE) on LAWS, where consensus has been stalled over definitions, human-machine interaction, and whether regulation should be voluntary or treaty-based.

International humanitarian law (IHL) remains central, with discussions focusing on how principles like distinction, proportionality, and precaution can apply to autonomous systems. Mohan also emphasizes a “life-cycle approach” to weapon assessment, extending legal and ethical oversight from design to deployment and decommissioning.

A significant portion of the conversation turns to gender and bias, an area Mohan has advanced through her research at UNIDIR. She draws attention to how gendered and racial biases encoded in AI systems can manifest in conflict, stressing the importance of diversifying participation in both technology design and disarmament diplomacy.

Looking forward, Mohan cites UN Secretary-General António Guterres’s call for a legally binding instrument on autonomous weapons by 2026. She argues that progress will depend on multi-stakeholder engagement, national strategies on AI, and confidence-building measures between states. The episode closes with a reflection on the future of warfare as inseparable from governance innovation—shifting from arms reduction to resilience, capacity-building, and responsible innovation.

Episode Contributors

Shimona Mohan is an associate researcher on Gender & Disarmament and Security & Technology at UNIDIR in Geneva, Switzerland. She was named among Women in AI Ethics’ “100 Brilliant Women in AI Ethics for 2024.” Her areas of focus include the multifarious intersections of security, emerging technologies (in particular AI and cybersecurity), gender, and disarmament. 

Charukeshi Bhatt is a research analyst at Carnegie India, where her work focuses on the intersection of emerging technologies and international security. Her current research explores how advancements in technologies such as AI are shaping global disarmament frameworks and security norms.

Readings

Gender and Lethal Autonomous Weapons Systems, UNIDIR Factsheet 

Political Declaration on Responsible Military Use of AI and Autonomy, US Department of State

AI in the Military Domain: A Briefing Note for States by Giacomo Persi Paoli and Yasmin Afina

Understanding the Global Debate on Lethal Autonomous Weapons Systems: An Indian Perspective by Charukeshi Bhatt and Tejas Bharadwaj