INDIA’S AI JOURNEY IN THE DEFENCE SECTOR: FROM EXPERIMENTATION TO DEPLOYMENT
Author: Dr. Binod Kumar Singh
Publication: The Calcutta Journal of Global Affairs, Volume 10, Number 3 (September 2026), © 2026
Publisher: Research Centre for Eastern and North Eastern Regional Studies, Kolkata (CENERS-K)
INTRODUCTION
Artificial Intelligence (AI) is a branch of computer science that focuses on developing systems that display aspects of intelligence: machines that acquire new concepts and skills, systems capable of reasoning and drawing practical inferences about their surroundings, programs that can comprehend natural language or interpret visual scenes, and other systems that perform tasks traditionally associated with human intelligence. [1] AI is a relatively new field that gained popularity in the 1940s and 1950s with the development of modern computers. The ability of these early computers to store enormous amounts of data and process it rapidly inspired researchers to conceive machines that could mimic certain human faculties. In 1956, John McCarthy, widely regarded as the father of AI, and his collaborators coined the term Artificial Intelligence. [2] The discipline emerged alongside the development of digital computers and is the result of centuries of efforts to build sophisticated machinery, coupled with advances in philosophy, mathematics, and logic within the scientific community. [3]
Historically, Japan was among the first nations to commit large-scale resources to AI research. In October 1981, it launched an ambitious ten-year program known as the Fifth-Generation Computer Systems (FGCS) Project, aimed at developing intelligent supercomputers. [4] Subsequently, other leading countries announced their own AI initiatives: the United Kingdom launched the Alvey Project, European nations cooperated through the European Strategic Programme on Research in Information Technology (ESPRIT) [5] program, and France pursued its own plan. In the United States, a private-sector group, the Microelectronics and Computer Technology Corporation (MCC), was formed in 1983 to advance technologies leveraging AI. Additionally, the Defense Advanced Research Projects Agency (DARPA) [6] increased support for AI research, funding three notable programs: Autonomous Land Vehicles (ALV), the Pilot’s Associate (PA) expert system assisting fighter pilots, and the Strategic Computing Program (SCP) aimed at creating AI-enabled military supercomputers. [7]
Recent advances in AI represent a major leap affecting many sectors globally. A Brookings report identifies AI—after the steam engine, electric power, and information technology—as the latest general-purpose technology. [8] The Internet of Things (IoT), smart sensors, and data-storing and processing devices, along with AI, are driving the ongoing development and commercialisation of new goods and services in what is commonly referred to as the Fourth Industrial Revolution. Prevailing goods and services are also finding new horizons with the integration of AI and IoT. [9] Gartner estimates that global AI spending reached $1.75 trillion in 2025 and projects $2.52 trillion for 2026 and $3.33 trillion for 2027. [10]
Table 1: World AI Spending by Market, 2025–2027 (Summary)
-
2025: $1.75 Trillion (1,750,000 USD Millions)
-
2026 (Projected): $2.52 Trillion (2,520,000 USD Millions)
-
2027 (Projected): $3.33 Trillion (3,330,000 USD Millions)
Source: Gartner, January 2026 [11]
(Note: Insert original Table 1 image from journal here if needed)
Meanwhile, the global AI market is expanding at a rapid pace. A United Nations Trade and Development (UNCTAD) report projects growth from $189 billion in 2023 to $4.8 trillion by 2033, positioning AI as the defining technology of the era. By 2033, AI has the potential to become the dominant force in the frontier technology market, when its market share could increase from 7 percent to 29 percent. [12] However, this fast growth could make global differences even bigger: development is mostly happening in large, advanced economies that have more skilled workers.
Figure 1: Estimated Frontier Technology Market Size and Share, 2023 and 2033
-
2023 AI Market Size & Share: $189 Billion (7% of frontier technology market)
-
2033 Projected AI Market Size & Share: $4.8 Trillion (29% of frontier technology market)
Source: United Nations Trade and Development (UNCTAD) [13]
(Note: Insert original Figure 1 chart from journal here if needed)
In 2022, only 100 companies, most of which were based in the US and China, conducted 40 percent of the world’s Research and Development (R&D) in AI. These two countries own 60 percent of AI patents and write a third of all AI articles. To stay competitive in an age of AI, developing countries need to rethink their industrial policies to put more emphasis on technology, innovation, and knowledge-intensive services. They also need to speed up their national AI strategies. UNCTAD recommends focusing on three key areas—infrastructure, data, and skills. Priorities should include:
-
Improving Infrastructure: Ensuring universal access to electricity, internet connectivity, and computational resources.
-
Encouraging Open Data: Promoting open data and data sharing to improve storage, access, and collaboration.
-
Building Societal AI Literacy: Ensuring that everyone in society is knowledgeable about AI by including Science, Technology, Engineering, and Mathematics (STEM) and AI education in schools and continuing education throughout life. [14]
Concurrently, AI-enabled systems—drones, autonomous platforms, and potentially lethal autonomous weapons—are creating complex legal questions in the absence of robust norms and regulations. This trend cuts across domestic and international legal frameworks. International law faces new challenges arising from rapid advances in AI programs, algorithms, and machines that increasingly influence global and local societies. Therefore, identifying the legal consequences of AI’s swift adoption and examining existing international law norms through the lens of this technology is crucial. [15] Indeed, AI is changing the military field by improving intelligence, surveillance, and reconnaissance (ISR), enabling drones and other autonomous systems, enhancing logistics and maintenance, and supporting command and control through real-time analytics that speed up and refine decisions—from border security to precision targeting—while also raising ethical concerns regarding autonomous weapons and cybersecurity.
With this context, this article examines how AI is changing military operations to boost efficiency, enhance decision-making, and improve the safety and efficacy of personnel. Its primary contribution is an analysis of India’s AI trajectory in defence, highlighting major AI projects in the field. The paper also offers a comparative overview of the global AI ecosystem to locate India’s standing. Furthermore, it discusses operational, ethical, and legal impediments to military AI adoption, proposing frameworks and measures to address them. Finally, it outlines strategic directions to strengthen AI in India’s defence sector and highlights future trends.
MILITARY APPLICATIONS OF AI
AI is applied across a wide spectrum of military tasks—from operational planning to medical response—improving efficiency and minimising human error. Key military uses of AI include:
-
Warfare Systems: Enhances weapon systems, sensors, navigation, flight support, and surveillance systems, increasing autonomy and operational efficiency, reducing reliance on human inputs, and lowering error rates while freeing personnel for higher-value roles.
-
Drone Swarms: Facilitates swarm behaviour among drones, enabling collaborative information sharing that strengthens reconnaissance and surveillance missions.
-
Strategic Decision-Making: Ingests and synthesises multi-source data in order to assist commanders, especially under stress. It provides rapid analytical support to supplement human ethical judgment and accelerates decisions.
-
Data Processing and Research: Rapidly filters and categorises diverse datasets, aiding pattern detection and strategy formation. This rationalises research workflows, enabling analysts to derive insights more efficiently and with fewer errors.
-
Combat Simulation: AI-driven military simulators or virtual wargames train personnel for real-world scenarios, bolstering skills while saving both time and resources compared to conventional training.
-
Target Recognition: Improves accuracy in target identification and classification, integrating multiple information sources to anticipate adversary behaviour and inform mission planning.
-
Threat Monitoring: Enhances situational awareness and threat detection by examining data collected from unmanned systems and drones, strengthening border security and protecting forces on the ground.
-
Cybersecurity: Defends military networks by detecting and responding to cyber threats, modelling attack patterns, and devising countermeasures to protect classified data and vital infrastructure.
-
Transportation: Enhances the movement of troops, supplies, and ammunition by charting optimal routes and forecasting disruptions, thereby cutting costs and improving logistics.
-
Casualty Care and Evacuation: Supports paramedics by analysing medical data and recommending treatments, improving decisions in high-pressure scenarios. [16]
In sum, AI reshapes military operations by raising efficiency, refining decisions, and improving personnel safety and mission effectiveness. It augments human capabilities, helping armed forces adapt more swiftly to evolving threats. Policymakers and analysts increasingly view AI as a transformative influence on future warfare, potentially altering the international balance of power. Yet, despite public interest in AI’s battlefield prospects, research is still nascent regarding exactly how AI will alter warfare.
To conceptualise AI’s role, one can separate its impact on warfighting from its role in preparing for war. Warfighting comprises active combat—dynamic and unpredictable, marked by friction and the fog of war—where adaptability is essential. AI might affect warfighting by:
-
Processing information and producing actionable knowledge;
-
Managing complex actions and systems;
-
Substituting human labour in certain tasks; and
-
Enhancing decision quality. [17]
By contrast, preparing for war includes force development, technological innovation, production, and logistics—functions relevant in peacetime and during extended conflicts. AI could influence preparation by improving production efficiency and enabling the design of novel military capabilities. [18] Combining autonomy with AI generally amplifies a weapon system’s effectiveness. The surge of interest in intelligent autonomous systems is reflected in China’s ambition to lead the AI field and the U.S. revival of such technologies through its Third Offset Strategy. [19] Central debates about autonomy and Machine Learning (ML) in weapons concern definitions and implications. ‘Autonomy’ typically refers to a machine’s capacity to carry out intended tasks without human intervention by leveraging sensors and software. ‘Machine Learning’ is an approach where systems are designed to learn from data and taught to perform tasks via various methods. [20] Both fall under the AI umbrella and are often used together.
INDIA’S AI JOURNEY IN THE DEFENCE SECTOR
India initiated its defence AI efforts in March 2018 by forming a Task Force on AI under the Ministry of Defence (MoD). In June 2018, the task force submitted its report with several important recommendations. In February 2019, based on the recommendations of the task force, the Defence Artificial Intelligence Council (DAIC) and the Defence AI Project Agency (DAIPA) were established. [21] In August 2019, an AI roadmap for Defence Public Sector Undertakings (DPSUs) was released. The MoD showcased its adoption of AI at the ‘AI in Defence’ (AIDef) symposium and exhibition in 2022. There, Defence Minister Rajnath Singh unveiled 75 products developed with the help of DPSUs, private companies, universities, and research organisations. The products included AI platform automation, C4ISR solutions, unmanned and autonomous robotic systems, voice and speech transcription and analysis systems, and an AI-enabled intrusion detection system. [22]
Although India’s national AI strategy initially emphasised a civilian, socio-economic orientation—leveraging AI for economic growth and inclusive development—the country is now channelling R&D efforts to foster an AI-enabled military ecosystem. [23] The MoD’s AI Task Force, established in March 2018, recommended in its June 2018 report that India should develop its AI power in defence in a significant way, particularly across land systems, naval systems, aviation, cyber, and nuclear and biological domains. [24] Later, in March 2024, India launched its ‘IndiaAI Mission’ to establish the country as a world leader in AI. The Mission comprises seven pillars:
-
IndiaAI Compute Pillar: Offers access to sophisticated GPUs at lower prices; more than 38,000 GPUs have been added and are available for ₹65 per hour.
-
IndiaAI Application Development Initiative: Concentrates on building AI applications that can help India address challenges in healthcare, farming, climate change, governance, and assistive learning. By July 2025, 30 applications had been approved. Sector-specific hackathons, like the Cyber Guard AI Hackathon, worked with ministries and institutions to develop solutions in areas like cybersecurity.
-
AIKosh: Collects large datasets from both government and non-government sources to train models. It houses a collection of over 3,000 datasets and 243 AI models covering 20 different sectors, helping developers focus on building solutions. By July 2025, the platform had recorded over 265,000 visitors, 6,000 registered users, and 13,000 downloads of datasets and models.
-
IndiaAI Foundation Models: Constructs India’s own large multimodal models trained on Indian data and languages to ensure the country possesses sovereign generative AI capabilities. IndiaAI received more than 500 proposals, out of which four startups—Sarvam AI, Soket AI, Gnani AI, and Gan AI—were chosen in the first phase.
-
IndiaAI Future Skills: Aims to develop AI talent by supporting 8,000 Undergraduate students, 5,000 Postgraduate students, and 500 Ph.D. scholars. By July 2025, more than 200 fellowships had been awarded, and 26 institutes began enrolling Ph.D. scholars. Data and AI labs have been set up in selected Tier-2 and Tier-3 cities. A total of 27 labs were identified with NIELIT, along with 174 ITIs and Polytechnics nominated by the States and UTs.
-
IndiaAI Startup Financing: Offers funding provisions for AI startups. The IndiaAI Startups Global program, launched in March 2025, has helped 10 Indian startups scale in Europe with the assistance of Station F and HEC Paris.
-
Safe and Trusted AI: Promotes accountable AI with governance mechanisms. In the first round, eight projects were selected focusing on machine unlearning, bias mitigation, privacy-preserving ML, explainability, auditing, and governance testing. In the second round, more than 400 applications were received. To join the IndiaAI Safety Institute, partner institutions were invited to submit an expression of interest on 9 May 2025. [25]
Considerable progress has already occurred in civilian AI deployment, and numerous Indian firms have built significant expertise that can be transferred to defence applications with relatively modest effort. [26] The Defence Research and Development Organisation (DRDO) is actively embedding AI and ML to modernise the armed forces, prioritising autonomous systems, robotic surveillance, and intelligent decision-support tools. DRDO laboratories such as the Centre for Artificial Intelligence and Robotics (CAIR) are working on AI-enabled UAVs, humanoid robots, and secure indigenous defence software aimed at delivering battlefield advantages by 2027. [27]
A notable CAIR accomplishment is the AI Techniques for Net-Centric Operations (AINCO) project—a specialised technology group for knowledge-base creation, semantic information handling, inference reasoning, and event correlation. [28] Additional CAIR projects comprise the Multi-Agent Robotics Framework (MARF), featuring systems like Wall-Climber robots and Snake robots, as well as research into image recognition for target identification and classification. [29]
India is accelerating AI adoption in defence to boost operational efficiency, surveillance, and decision-making, aiming for self-reliance in future warfare. Under DAIC guidance, initiatives encompass AI-enabled border surveillance, autonomous drones, and predictive maintenance. In September 2025, India released a 15-year defence roadmap (2025–2040) focused on AI, hypersonics, and next-generation warfare—charting a major modernisation path to address emerging challenges over the coming decades. [30]
By February 2026, the MoD had signed 548 contracts under the Innovations for Defence Excellence (iDEX) initiative covering AI, autonomy, and other advanced technologies, engaging more than 676 startups and MSMEs. [31] A total of 106 projects using AI are now recognised, while 40 tools built on AI technology have entered defence operations. [32] India is pursuing a multi-layered security approach, prioritising 75 projects centred on AI, drone swarms, and cybersecurity, emphasising indigenisation as part of the Atmanirbhar Bharat (Self-Reliant India) [33] drive—especially to bolster border security along the Pakistan and Bangladesh frontiers. [34]
Notably, the Indian Army made significant use of AI during Operation Sindoor (May 7–10, 2025) against Pakistan, generating a unified operational picture, conducting intelligence analysis, developing threat assessments, generating predictive modelling, and performing weather forecasting for long-range strikes. Overall, 23 applications with specific functions were employed to process data and inputs. Key AI-driven tools included the Electronic Intelligence Collation and Analysis System (ECAS)—domestically developed and upgraded to real-time capability—to identify and prioritise grave threats, helping secure strategic advantages. [35]
INDIA’S KEY DEFENCE AI PROJECTS
India’s principal defence AI efforts concentrate on autonomous surveillance, intelligence analysis, and combat support under DAIC leadership. Notable projects include Project Storm Drone, AI-equipped ROVs, and the SAM-UN situational awareness system—all intended to modernise the armed forces via automation, AI-powered speech recognition, and resilient intelligent networks. India’s key projects include:
-
Surveillance & Intelligence: AI-enabled surveillance robots for borders, drone-based room intervention (Project Storm Drone), and AI-capable Remotely Operated Vehicles (ROVs) for underwater inspection.
-
Situational Awareness & Mapping: SAM-UN, a geospatial platform for situational awareness, mission planning, and disaster response.
-
Battlefield & Combat Tools: AI applications for speech recognition, image processing, and the ECAS for precision targeting.
-
Predictive Logistics & Analysis: Brahapar (military climatology for flood and avalanche forecasting) and Rakshapan (predictive logistics modelling).
-
Cybersecurity & Recognition: Advanced deepfake detection systems, AI-enabled cybersecurity, and XFace for facial recognition.
-
Defence AI Ecosystem: Through DAIC and DAIPA, India nurtures AI startups via iDEX and allocates ₹100 crore annually. [36]
As India prepares to celebrate its centenary of independence in 2047, the Viksit Bharat (Developed India) [37] vision sets an ambitious, inclusive path to developed-nation status. Achieving that requires immediate action. India’s digital public infrastructure has previously proven its success, as seen through Aadhaar, UPI, and Jan Dhan, demonstrating the country’s capacity to build large-scale, inclusive platforms. With a robust digital backbone, a youthful workforce, and an untapped informal sector, India has a rare opportunity; delay risks leaving millions behind and impeding development. Bold, purposeful action can place informal workers at the heart of growth and secure the foundation of a Viksit Bharat. [38]
Major DPSUs Leading AI Integration in India’s Defence
-
Bharat Electronics Limited (BEL): Has taken the lead in implementing AI in collaboration with the Indian Army through an AI Incubation Centre. Some of its innovative solutions include the use of AI-enabled Voice Analysis Software (AIVAS) to assist with transcription and language analysis; a document management system called Sandarbh.AI for the Indian Navy; and a predictive maintenance system to improve fire control systems. [39]
-
Hindustan Aeronautics Limited (HAL): Through its integration of AI, aerospace systems such as the Combat Air Teaming System (CATS) utilise AI technology to allow drones to respond to voice commands issued by fighter pilots. Additionally, automated defect identification can be accomplished using an AI-enabled Snag Disposition System, and maintenance and carbon-detection tools can be utilised for aircraft components using AI. [40]
-
Bharat Earth Movers Limited (BEML): Develops AI-enabled systems for monitoring and logistics, such as the Driver Fatigue Monitoring System. It is advancing AI integration in India’s defence sector by partnering with institutions like the Defence Institute of Advanced Technology (DIAT) to modernise technologies and develop autonomous, high-mobility vehicles. [41]
-
Garden Reach Shipbuilders & Engineers (GRSE): Focuses on the evaluation of welding defects with AI-enabled systems in Non-Destructive Testing (NDT) and AI integration in naval platforms. Through ‘GAINS 2025,’ the shipyard uses AI for surveillance and design optimisation to support the Indian Navy and Coast Guard. [42]
Private Sector Companies Leading AI Integration in India’s Defence
-
Data Patterns (India) Limited: Integrates AI with radar and avionics technologies, improving object detection capabilities for surveillance purposes and working with DRDO as well as ISRO to develop next-generation AI-based systems. [43]
-
Zen Technologies Limited: One of the first companies in India to offer an anti-drone system that uses ML to detect and disable drones; their military simulation systems utilise AI to provide realistic training for the armed forces, paramilitary forces, and police for actual battlefields. [44]
-
Paras Defence and Space Technologies: Has invested in Logic Fruit Technologies for advanced AI-enabled embedded solutions. It has also developed PARAS.AI for drone data analytics and is developing AI-enabled radar and electronic warfare systems. [45]
-
MTAR Technologies: Currently developing AI-enabled manufacturing processes and collaborating with global defence partners to manufacture and develop mission-critical, precision-engineered systems for the aerospace and defence industries. [46]
In order to mainstream AI across all military operations, Defence Minister Rajnath Singh launched ‘Defence Forces Vision 2047: A Roadmap for a Future-Ready Indian Military’ on 10 March 2026. This document presents a long-term plan for modernising the Indian military into a cohesive and technology-savvy force by 2047, marking the centenary of Indian independence. Produced by the Headquarters Integrated Defence Staff, the document presents a blueprint for strengthening competencies aligned with the Viksit Bharat goal. [47] DAIPA functions as the implementation arm tasked with converting the policy into deployed technologies.
In a notable modernisation effort, the Indian Army plans to convert its ageing T-72 main battle tanks into remotely operated and autonomous armoured fighting vehicles to extend their operational life beyond their planned retirement starting in 2030. The initiative intends to adapt legacy platforms to technology-driven warfare, enabling these tanks—of which roughly 2,400 remain in service—to fulfil roles like minefield breaching, decoy deployment, advance guard duties, and reconnaissance without onboard crews, thereby reducing casualties while preserving capability. [48]
From its use in military education at institutions such as the National Defence Academy (NDA) to its application in wargaming, simulations, and network-centric warfare, AI has become integral to the Indian armed forces’ training and conduct of war. What was once considered futuristic has now become ubiquitous in training, planning, and warfighting, signifying a move towards data-centric and learning-based warfare. Recent conflicts and operations, such as Operation Sindoor, have been a manifestation of such complex factors. [49]
COMPARATIVE ANALYSIS OF THE GLOBAL AI ECOSYSTEM
In the 2025 Global AI Index of Stanford University, India ranks as the world’s third-most dynamic AI ecosystem, behind the United States and China. [50] India’s position stems from a strong talent base, a growing startup ecosystem, and rising AI research and patent activity. However, despite strengths in talent, startups, and R&D, India remains a ‘distant third,’ substantially trailing the top two in areas such as infrastructure and investment. The gap between India and major military powers is stark. In 2025, the United States led global AI investment, followed by China and the UK, with these nations dominating public and private funding concentrated on AI infrastructure and compute. [51]
Table 2: Top 10 Countries by AI Investment, 2013–2025
Led by the United States, China, and the United Kingdom in total public and private AI infrastructure and compute investment.
Source: Forbes [52]
(Note: Insert original Table 2 image from journal here if needed)
Beyond funding disparities, effective AI requires vast datasets and compute infrastructure for training. TRG Datacenters reports that the United States was the most AI-dominant country in 2025, with total AI compute power equivalent to 39.7M H100 units and a power capacity of 19.8K MW—indicating substantial energy resources for AI. The UAE ranks second with 23.1M H100 equivalents and 6.4K MW capacity; Saudi Arabia follows with 7.2M compute equivalents and 2.4K MW capacity. South Korea stands fourth with 5.1M equivalents and 3K MW. France is fifth at 2.4M compute. India occupies the sixth position with 1.2M H100 equivalents, 493K AI chips (third highest), and a total of 1.1K MW power capacity. China is in the seventh position—despite having the most clusters (230)—with 400K H100 equivalents and 289 MW total capacity, while holding 629K AI chips. The UK, Finland, and Germany round out the top ten with significantly lower compute and power figures. [53] Practically, this means India trains military AI systems with markedly smaller compute and dataset capacities than its primary geopolitical rivals.
Table 3: World’s Top 10 AI Supercomputer Powers in 2025 (Key Metrics)
| Rank |
Country |
Compute Power (H100 Equivalents) |
Power Capacity |
Notable Hardware / Clusters |
| 1 |
United States |
39.7M |
19.8K MW |
Global leader in compute & power |
| 2 |
United Arab Emirates |
23.1M |
6.4K MW |
Major hyperscale AI hub |
| 3 |
Saudi Arabia |
7.2M |
2.4K MW |
Rapidly expanding sovereign compute |
| 4 |
South Korea |
5.1M |
3.0K MW |
High density of AI compute |
| 5 |
France |
2.4M |
— |
Leading European AI compute power |
| 6 |
India |
1.2M |
1.1K MW |
493K AI chips |
| 7 |
China |
400K |
289 MW |
230 clusters; 629K AI chips |
| 8–10 |
UK, Finland, Germany |
Lower compute figures |
Lower power figures |
— |
Source: TRG Datacenters [54]
In the race to build military AI, data has become a decisive asset—with superiority increasingly measured in algorithms as much as in conventional armaments. Possession of data currently confers a significant advantage in AI development. However, as militaries rely more on AI, data centres have emerged as strategic targets. During the early phases of the ongoing United States–Israel–Iran conflict, data centres were struck. On 8 March 2026, Iran hit two Amazon Web Services (AWS) data centres in the UAE and one in Bahrain, causing moderate damage but widespread disruption across banking and consumer services. [55] Conversely, on 11 March 2026, a strike affected a data centre tied to Bank Sepah in Tehran, disrupting salary payments for Iranian military units. [56]
On the same day, an IRGC-affiliated outlet published a list of 29 “tech targets” across Bahrain, Israel, Qatar, and the UAE, including facilities belonging to IBM, AWS, Microsoft, Google, Palantir, Nvidia, and Oracle. [57] By attacking Gulf data centres, Iran has raised global questions about the resilience, sovereignty, and security of digital infrastructure—blurring the line between commercial cloud services and military operations. Critical military systems such as the Pentagon’s ‘Joint Warfighting Cloud Capability’ and ‘Joint All-Domain Command and Control’ operate on the very same infrastructure as banks and consumer services. There are reports suggesting the United States military utilised Anthropic’s Claude, operating on AWS infrastructure, for intelligence assessments as well as battle simulations in strikes against Iran. The implications of strikes on such infrastructure are immediate. [58] Data centres are attractive military targets because they are large in size and reliant on vulnerable support infrastructure such as cooling units, generators, and turbines, which can be easily identified via satellite imaging. They are also located close to reliable energy and water resources, making their locations easy to pinpoint despite the secrecy involved, while their close proximity to urban areas leaves civilian populations vulnerable. [59]
CHALLENGES IN THE USE OF AI IN DEFENCE
Implementing AI in defence provides significant benefits, but it also introduces multiple challenges spanning technical, ethical, legal, and operational domains:
-
Technical Challenges: Ensuring data security, addressing algorithmic bias, meeting heavy computational demands, and maintaining system reliability in unpredictable combat environments. [60]
-
Ethical Challenges: The erosion of human oversight and accountability in lethal decisions, biased targeting outcomes, and difficulties assigning responsibility for potential war crimes. [61]
-
Legal Challenges: Accountability for autonomous actions, adherence to International Humanitarian Law (IHL), and biases embedded in algorithms. [62]
-
Operational Challenges: Limited transparency, scarcity of labelled training data, algorithmic bias, and integration issues with legacy systems. [63]
A notable operational risk is AI accelerating the pace of warfare beyond human ability to respond—a phenomenon termed ‘battlefield singularity’ or ‘hyper-war,’ which could cause strategic errors, accidents, or unintended escalation. Even if some risks are mitigated, increasing AI reliance may diminish human factors in military strategy—such as psychology and judgment—potentially creating a divergence between how AI solves a problem and how humans would act with equivalent speed and precision. [64]
Joint military operations involving multiple actors present additional complexities. While current joint operations require doctrinal and procedural alignment, AI introduces further requirements for interoperability. Variations in AI acceptance, application domains, data requirements, algorithmic design (producing different outputs), and analytical recommendations raise questions about harmonisation and mutual acceptance of AI tools among coalition partners. Addressing these challenges demands coordinated action across governments, industry, academia, and international organisations to ensure responsible and effective defence AI deployment. [65]
Given AI’s central role in future decision-making, strategy documents promoting AI must proactively address ecosystem risks that could compromise ethics, privacy, or security. Risk mitigation measures should be integral to any AI policy. [66] Several countries have adopted international principles on the use of AI in the military domain by committing to responsible frameworks, including:
-
The Political Declaration on the Responsible Military Use of AI and Autonomy, led by the US (2023);
-
The Blueprint for Action by the Netherlands and South Korea at the Responsible AI in the Military Domain (REAIM) Summit (2024); and
-
The Paris Declaration on Maintaining Human Control in AI-Enabled Weapon Systems at the AI Action Summit (2025).
This involves implementing AI principles by ensuring military doctrine, tactics, techniques, and procedures are compatible with AI technologies, as well as ensuring the acquisition and use of AI technologies remain consistent with responsible norms. [67]
STRATEGIES TO IMPROVE AI IN INDIAN DEFENCE
Strengthening AI within Indian defence calls for boosting indigenous R&D, enhancing public–private partnerships, and fortifying data and digital infrastructure to create secure, dependable, and intelligent systems. Priority deployments include AI-based surveillance, swarm drones, and predictive logistics, supported by an empowered DAIPA and ethical AI frameworks. The key strategies to improve AI in Indian Defence are as follows:
-
Foster Indigenous Innovation & Partnerships: Bridge gaps between DRDO, DPSUs (HAL, BEL, BDL), and private tech firms to enable agile development. Scale up funding for startups via iDEX to accelerate solutions in AI, sensor systems, and autonomous unmanned platforms. [68]
-
Strengthen Data & Digital Infrastructure: Build a secure national repository of labelled, defence-relevant data to train algorithms while ensuring data sovereignty. Invest in domestic high-performance computing centres and indigenously designed semiconductors for sensitive defence workloads. [69]
-
Focus Areas for Deployment: Deploy AI-based image-analytics systems (e.g., Agni-D) for real-time border monitoring and threat detection. Develop AI-enabled swarm drones for collaborative surveillance and target identification. Apply AI to historical datasets for demand forecasting, supply-chain optimisation, and preventive maintenance. [70]
-
Institutional & Human Resource Development: Empower DAIPA to set AI development standards, Standard Operating Procedures (SOPs), and Intellectual Property Rights (IPR) policies. Introduce specialised AI courses across defence training institutions. [71]
-
Ethical AI Frameworks: Maintain human-in-the-loop (HITL) architectures for critical life-or-death decisions. Through initiatives such as M.A.N.A.V., India aims to promote an AI trajectory that advances both technology and human welfare. [72]
-
Strategic Initiatives: Improve military decision-making with wargaming and simulation at centres like the Wargame Research and Development Centre (WARDEC), recognising wargaming’s growing role in operational planning, leadership development, and doctrinal innovation for multi-domain operations. [73]
Given India’s inability to match the dataset scale of the United States and China, its defence sector and deep-tech startups are pivoting towards smaller language models (SLMs) and edge AI to create smarter battlefield systems. Long-term strategy should emphasise indigenous AI models attuned to domestic defence datasets. Sarvam AI is a notable example selected under the IndiaAI Mission’s Innovation Centre pillar to build foundational models for Indian languages and public services, offering voice interfaces, document processing, and citizen-facing applications that reduce reliance on foreign systems and nurture an open-source ecosystem spanning startups, academia, research institutions, and industry. [74]
Of late, India has attracted substantial investments from global tech players into its AI ecosystem. For example, on 7 January 2025, Microsoft pledged a $3 billion investment to build data centres in Telangana. [75] Separately, the United States–India Taskforce on Artificial Intelligence (launched in early 2025) seeks to accelerate bilateral defence collaboration, enhance AI infrastructure, support research, and establish ethical standards for military AI use. Its objectives include joint AI tool development, startup support, and integrating AI into surveillance and autonomous systems. [76] Moreover, at the AI Impact Summit 2026 (New Delhi, February 2026), Indian firms and the Government pledged investments exceeding $250 billion in AI infrastructure to democratise AI access. [77]
CONCLUSION
AI’s introduction into military affairs marks a fundamental shift in defence strategy. As AI capabilities accelerate task execution, boost productivity, and lower human error, nations see a strategic opportunity in harnessing these strengths. Yet embedding AI into military contexts presents technological, ethical, legal, and operational challenges. For trustworthy AI systems, there is a need for large amounts of quality data, alongside concerns over the amplification of existing vulnerabilities, the creation of new security risks, and the escalation of conflicts beyond human control. Ethical considerations, particularly those concerning life-and-death choices, continue to dominate.
Tackling these issues requires a comprehensive approach uniting government, industry, academia, and international institutions. Strong regulation, continuous oversight, and governance frameworks are essential to ensure ethical and effective military AI usage. As a new entrant and a distant third in global rankings, India must also scale computational capacity and energy provision, as power constitutes the largest operating cost for hyperscale AI data centres. Additional gaps include scalable deployment, interoperability across Army, Navy, and Air Force systems, and access to defence-grade datasets. Without full integration among the three Services, AI models will not be able to cross-reference data efficiently, compromising a holistic operational picture. Moreover, predictive AI depends on large, unbiased datasets—an enduring challenge.
Despite promising advances, integrating AI into India’s defence structure remains complex. Data privacy, cybersecurity, and ethical issues around autonomous systems demand careful attention. Concerns about accountability—especially involving lethal autonomous weapons—persist. As global competition for AI leadership intensifies, India must keep pace through sustained R&D investment and international collaboration to exchange knowledge and best practices. Ultimately, India’s AI trajectory in defence is moving decisively from concept to operational reality—signalling a broader shift from experimentation to field deployment.
ENDNOTES & REFERENCES
-
Dan W. Patterson, Introduction to Artificial Intelligence and Expert Systems, Prentice-Hall Incorporated, New Jersey, 1990, p. 2.
-
Richard E. Neapolitan and Xia Jiang, Artificial Intelligence, CRC Press, London, 2018, p. 4.
-
Deepak Khemani, A First Course in Artificial Intelligence, McGraw Hill, Chennai, 2013, p. 2.
-
Edward Feigenbaum and Howard Shrobe, ‘The Japanese National Fifth Generation Project: Introduction, survey, and evaluation,’ Future Generation Computer Systems, Vol. 9, 1993, pp. 105–117.
-
ESPRIT was launched in the 1980s as a major European Commission initiative to promote collaborative research and development of IT amongst the companies and universities of the European Community.
-
DARPA was founded in 1958 as a research and development agency of the United States Department of Defense to develop emerging technologies for the United States military.
-
Dan W. Patterson, Introduction to Artificial Intelligence and Expert Systems, p. 4, n. 1.
-
-
-
-
Ibid.
-
-
Ibid.
-
Ibid.
-
Jaemin Lee, Artificial Intelligence and International Law, Springer, Singapore, 2022, p. 2.
-
Shalini Sharma, ‘The rise of artificial intelligence in military: A global perspective,’ International Journal of Political Science and Governance, Vol. 7, No. 6, 2025, pp. 11–16.
-
-
-
-
-
-
-
-
-
-
-
-
-
Petr Topychkanov, ‘The Impact of Artificial Intelligence on Strategic Stability and Nuclear Risk,’ op. cit., n. 19.
-
-
-
-
Atmanirbhar Bharat is a flagship policy framework and vision launched by Prime Minister Narendra Modi and the Government of India on May 12, 2020. The mission aims to make India self-sustaining, resilient, and less dependent on imports by boosting domestic manufacturing, upgrading infrastructure, and fostering technological innovation across sectors.
-
-
-
-
Viksit Bharat is the Government of India’s ambitious vision to transform India into a fully developed nation by the 100th anniversary of its independence in 2047. It aims for economic prosperity, social progress, environmental sustainability, and good governance, focusing on four pillars: youth, the poor, women, and farmers.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Ibid.
-
Ibid.
-
-
-
-
-
-
-
Pooja Awasthi, ‘Artificial Intelligence in Future Defence Systems: Opportunities, Risks, and Ethical Challenges,’ Journal of Applied Bioanalysis, Vol. 11, No. 11, December 2025, pp. 1079–1087.
-
-
Amir Rabet Ghanna, ‘Artificial Intelligence and International Law: Challenges and Opportunities,’ Legal Studies in Digital Age, Volume 5, Issue 1, January 2026, pp. 1–15.
-
-
-
Shalini Sharma, ‘The rise of artificial intelligence in military: A global perspective,’ International Journal of Political Science and Governance, Vol. 7, No. 6, 2025, pp. 11–16.
-
National Strategy for Artificial Intelligence, NITI Aayog, op. cit., n. 23.
-
-
Sudhish Kumar Jaini and Navin Varma, ‘Indian defence manufacturing landscape—DRDO, public & private sector,’ International Journal of Applied Research, Vol. 11, No. 10, 2025, pp. 179–197.
-
-
-
Pardeep Bhardwaj, ‘Use of Artificial Intelligence for Improving Operational Efficiency in the Armed Forces,’ Air Power Journal, Vol. 19, No. 4, Winter 2024 (October–December), pp. 1–23.
-
-
-
-
-
-
ABOUT THE AUTHOR
-
Dr. Binod Kumar Singh is presently Guest Faculty at the Department of National Security Studies, Manipur University, Manipur. After completing his Ph.D. in 2012 from the Department of Defence and National Security Studies, Panjab University, Chandigarh, he joined the Institute for Conflict Management (ICM), New Delhi, and worked as a Research Associate until April 2022. Over the past 14 years, he has published nine articles in refereed journals and nine chapters in edited books.