Defence forces move towards unified battlefield, ISR & cyber data: DG DRDO

What should be India’s long-term vision for AI-enabled defence capabilities? 

India’s long-term vision for AI-enabled defence should be to treat AI as strategic infrastructure that ensures deterrence, decision superiority and sovereign capability for our defence forces. We aim to be in the top-3 among global AI powers with resilient, human-overseen systems across multi-domain ops. From the application perspective, AI will serve as a prominent technology to enhance the defence capabilities of the soldier, systems and operations in the military domain in order to achieve better speed, accuracy and precision in highly contested battlefield environments. From a technology perspective, the indigenous defence AI is being developed around five core principles of predictability, reliability, traceability, resilience and trustworthiness.

How can DRDO define priority AI mission areas aligned with India’s national security doctrine? 

India’s National Security Strategy tenets include credible deterrence, border security, maritime domain awareness and multi-domain superiority. We know that in June 2018, the Government of India published the National Strategy for AI, focusing on how India can leverage the transformative AI technologies to ensure inclusive growth in all sectors of the society. And subsequently in February 2021 Government published principles for Responsible AI. In consonance with these policy initiatives of the Government, DRDO has also formulated guidelines and framework for evaluating trustworthiness in AI as well as IV&V (Independent Verification & Validation) Framework for AI-based systems released in year 2025 and 2026 respectively. 

DRDO has been actively working in niche application areas of AI for ISR applications, man-machine teaming, Natural Language Processing, Unmanned Autonomous Systems etc. Against asymmetric threats from adversaries like, for example, massed swarms and hybrid tactics, embedding of user-aligned and trustworthy AI, human oversight and edge compute efficiency are being incorporated at the design stage itself.

In addition to these DRDO will keep exploring new application areas of AI as per use cases. In upcoming times, DRDO will also focus towards promulgation and implementation of Trustworthiness & V&V (Verification & Validation) frameworks in conjunction with all the stakeholders including the armed forces, QA agencies, academia, industry and Department of Telecom of Standardisation. 

Should India pursue full-stack sovereign AI for defence, or selective strategic autonomy? 

India should pursue a hybrid full-stack sovereign AI for defence, controlling the critical stack (data, compute, core models, deployment) while selectively pursuing autonomy in high-risk missions like ISR, avoiding overreach that delays capabilities. The current AI applications are being developed with available tool-chains, data-sets and computation resources in order to meet the immediate requirements of the Services. Vulnerabilities are being addressed by rigorous verification and validation.

What is the update on multiple labs and centers, like the Centre for Artificial Intelligence & Robotics, DRDO has dedicated to AI? 

CAIR has been working extensively on fundamental and applied research in AI for defence-specific problems. This includes machine perception, AI-based decision support, reasoning systems and knowledge extraction which can be applied to unmanned systems, autonomous platforms and real-time information analysis for battlefield scenarios. In addition to this, as per directions of the Prime Minister, DRDO created a dedicated Young Scientist Laboratory in the area of AI to carry out research in the niche-technological domain. DYSL AI centre at Bangalore has been working on AI-based advanced technologies and language translation for situational awareness. 

What governance model should DRDO adopt for AI program management? 

DRDO is adopting a ‘Whole of the nation approach’ as a systemic model for governance of futuristic AI systems in defence applications. Under this model it is proposed that the Armed Forces institutionalise AI safety requirements as part of the acquisition life-cycle process, and academia should establish V&V maturity models for Edge-AI and explainability. Industry should develop indigenous AI V&V tool-chains (model probes, adversarial test-beds), and QA agencies should formalise AI V&V test labs and sandboxes for qualification and airworthiness, and DRDO can contribute towards building standardised defence AI solution suites for each domain based on user requirements as well as contribute towards building defence-grade secure data repositories for training, validation and testing.

Better coordination within the Armed Forces is also required for real-time operational feedback. How can that happen?

The Ministry of Defence (MoD) has constituted the Defence AI Council (DAIC) in February 2019 as the apex decision-making body for artificial intelligence adoption in India’s defence sector. It was established as part of India’s structured push toward AI-enabled military capabilities. DAIC provides strategic direction and policy-level oversight for AI in the Armed Forces. At the working level, Defence AI Project Agency (DAIPA) works as an executive arm that implements AI projects approved by DAIC. It operates under the MoD/DRDO framework. DAIPA translates AI strategy into operational projects and deployable systems. These endeavors will ensure that our Services will have the battlefield edge as they are part of the real-time feedback loops via field trials and the iterative process for the DRDO developed products.

What lessons can India draw from AI defence ecosystems in the United States and China? 

The United States operates with the model of institutional integration and China works with the model of state-driven military–civil fusion. The US created the Chief Digital & Artificial Intelligence Office, merging AI, data, analytics and digital transformation into one unified authority. In Indian context, DAIC (strategy) and DAIPA (projects) can integrate data governance, cloud, AI, and cyber under one operational umbrella; and ensure cross-service interoperability early. However, in case of China’s state-driven military-civil fusion strategy, civilian AI breakthroughs rapidly migrate into defence systems.

On the similar lines, India may strengthen structured collaboration between DRDO, IITs, IISc and private AI firms, incentivise dual-use AI research; and create classified research corridors in academia. In addition to this, India’s AI strategy will be executed in consonance with the PM’s vision of MANAV for artificial intelligence. MANAV, the hindi word for ‘human’, serves as an acronym representing five core principles, namely; Moral & Ethical System, Accountable Governance, National Sovereignty, Accessible & Inclusive Technology; and Valid/Legitimate systems.

There is also a need to standardise battlefield, ISR, UAV, naval and cyber data. Is there a move towards that? 

There is a growing recognition in Indian defence of the need to standardise and integrate data across battlefields, ISR (Intelligence, Surveillance & Reconnaissance), UAV, naval and cyber domains, and early steps toward that are underway. Modern military AI and decision-support systems rely on consistent, interoperable data formats and pipelines, especially when fusing data. DRDO’s ETAI Framework (Oct 2024) mandates pipeline standards for data annotation, versioning and secure handling across AI/ML artifacts.

Data localisation and protection from foreign cloud exposure is equally important. Your thoughts.

For defence AI, data localisation and insulation from foreign cloud exposure are strategic imperatives, not just IT preferences. In addition to privacy concerns, data localisation and protection is essential in order to ensure sovereignty, operational continuity and wartime resilience. Mandating sovereign infrastructure clusters for classified data with air-gapped networks has to be the norm. This is in line with our NSS. Government is planning to roll out a comprehensive five-layer AI Stack Strategy actively developing capabilities across Application, Model, Compute, Infrastructure & Energy. We strongly believe that this multi-layered development of indigenous AI stack will duly address challenges of data localisation and protection from foreign cloud exposure.