The Thinking Battlefield: Artificial Intelligence and India's Military

Every era of warfare has had its defining technology — the stirrup, gunpowder, the aircraft, the nuclear weapon. The defining technology of the coming era may be less visible than any of these, because it lives not in steel but in code. Artificial intelligence — the ability of machines to perceive, learn, decide and act — is beginning to seep into every corner of military affairs, from the drone that identifies its own targets to the software that sifts oceans of surveillance data to the systems that will one day help commanders decide.

India, home to one of the world's great pools of software talent and one of its largest militaries, has recognised that it cannot sit out this transformation. It is investing in military AI across surveillance, autonomy, cyber and decision support, while grappling with the profound ethical and strategic questions the technology raises. This is the story of how the thinking battlefield is arriving, what India is doing to be ready for it, and why the stakes are as much moral as military.

Why AI changes war

The power of artificial intelligence in warfare comes from a simple asymmetry: machines can process information at a scale and speed no human can match. Modern militaries are drowning in data — from satellites, drones, radars, sensors, intercepted communications and a hundred other sources. The bottleneck is no longer collecting information but making sense of it fast enough to act. AI attacks precisely this bottleneck. An algorithm can scan thousands of hours of surveillance footage, flag the one vehicle that matters, correlate it with other data, and present a commander with an answer in seconds rather than the days a human team would need.^[1]

From that core capability, the applications radiate outward. AI can guide autonomous vehicles on land, sea and air, letting machines navigate and operate without constant human control. It can sharpen precision, helping weapons and sensors identify and track targets. It can defend and attack in the cyber domain at machine speed, where human reflexes are far too slow. It can optimise logistics, predict when equipment will fail, and simulate battles to train commanders. And, most controversially, it can be built into weapons that select and engage targets with little or no human intervention — the domain of autonomous weapons that has become the central ethical battleground of military AI.

The bottleneck of modern war is no longer seeing the battlefield. It is understanding it fast enough to act.

The race India cannot skip

India's motivation is, as ever, partly the neighbourhood. China has declared its ambition to lead the world in artificial intelligence and has made military AI a national priority, investing heavily in autonomous systems, AI-enabled surveillance and algorithmic warfare. A rival that fields AI-enabled forces — faster to see, faster to decide, faster to strike — would enjoy a decisive advantage over an adversary still fighting at human speed. India cannot allow such a gap to open on its borders, and this strategic reality has pushed military AI from an interesting possibility to a national necessity.^[2]

India also has, unusually, a genuine comparative advantage to bring. It possesses one of the largest and most capable software and information-technology talent pools on earth, a thriving technology sector, and a young population fluent in the digital world. If any domain of military competition plays to India's strengths, it is one built on code and mathematics rather than on heavy industry. The challenge is to channel that civilian technological depth into military capability — to build the bridges between India's software prowess and its armed forces that turn national talent into national defence.

Building the institutions

Recognising that AI would not integrate itself, India moved to build the institutions to drive it. The government established a dedicated framework to promote artificial intelligence in defence, including a task force that studied how AI could be applied to the military and a body to oversee its development and adoption. Defence public-sector units and research laboratories were tasked with AI projects, and the private sector and start-ups were drawn in through challenges, funding and procurement aimed specifically at defence AI.^[3]

The institutional push reflects an understanding that military AI is not a single weapon to be procured but an enabling technology to be woven throughout the force — into surveillance, logistics, maintenance, cyber, training and command. Building it requires sustained investment, a pipeline of talent willing to work on defence problems, and mechanisms to bring the fast-moving civilian AI world into contact with the slower, more careful world of military procurement. India's institutions for this are young, but they exist, and they signal that the country intends to be a builder of military AI rather than merely a consumer of it.

Where India is applying it

The most immediate and least controversial applications are in surveillance and situational awareness. Along India's long and demanding borders, AI-enabled systems can help monitor vast and difficult terrain, analyse imagery from drones and satellites, detect intrusions, and flag the anomalies a human watcher might miss in an endless feed. In a country that must guard thousands of kilometres of frontier across mountains, deserts and coastlines, the force-multiplying effect of AI-assisted surveillance is enormous — it lets a finite number of human eyes cover an effectively infinite amount of watching.^[4]

Autonomy is the second great domain. India is developing unmanned systems across air, land and sea, and AI is the brain that makes true autonomy possible — the capacity for a machine to navigate, adapt and operate without a human controlling its every move. Coupled with the country's parallel push in drones and swarms, AI-enabled autonomy points toward forces in which machines take on the dullest, dirtiest and most dangerous tasks, extending the reach of human soldiers while keeping them out of the worst harm's way.

Then there is the invisible domain of cyber and electronic warfare, where AI is arguably most essential, because attacks and defences unfold at speeds no human can follow. And there is the quieter but hugely valuable application of predictive maintenance and logistics — using AI to forecast when an aircraft engine or a vehicle will fail, to optimise the flow of supplies, and to keep a vast military machine running more efficiently. These unglamorous uses may deliver more real value, sooner, than any autonomous weapon.

The ethical fault line

No honest account of military AI can avoid its darkest question: the prospect of weapons that kill without a human deciding. Lethal autonomous weapons systems — machines that can select and engage human targets on their own — sit at the centre of an intense global ethical and legal debate. Critics, including many scientists and human-rights advocates, warn that delegating the decision to take a human life to an algorithm crosses a moral line that should never be crossed, and that such weapons could lower the threshold for war, malfunction catastrophically, or proliferate to the worst actors.^[5]

India has generally positioned itself carefully in this debate, supporting the responsible development of military AI while engaging with international discussions on the norms that should govern autonomous weapons. The prevailing principle in most serious militaries, India's included, is that of meaningful human control — the insistence that a human being remains in or on the loop of any decision to use lethal force, that the machine advises and assists but the human decides. Holding to that principle as the technology grows more capable, and as rivals may be tempted to abandon it for the advantage of machine speed, will be one of the defining ethical challenges of the coming decades.

The dilemma is genuine and unresolved. Meaningful human control preserves moral responsibility and guards against catastrophic error, but it also imposes the limit of human reaction time in a domain where speed may decide survival. A nation that binds itself to human control while an adversary does not may find itself outpaced; a nation that abandons it may find itself having built something monstrous. There is no easy answer, and the tension between military necessity and moral restraint runs through the entire enterprise of military AI.

The hardest question of military AI is not whether the machine can decide to kill. It is whether we should ever let it.

The risks beyond ethics

Even setting aside the deepest moral questions, military AI carries serious practical risks. AI systems can be brittle — confident and wrong, fooled by inputs they were not trained for, or manipulated by an adversary who understands their weaknesses. An enemy who can deceive an AI surveillance system, poison the data it learned from, or spoof the sensors it relies on can turn a military's own algorithms against it. Building AI that is robust, trustworthy and secure against such manipulation is as important as building AI that is capable in the first place.^[1]

There is also the danger of over-reliance and of speed itself. A world in which decisions are made at machine pace is a world in which crises can escalate faster than human judgment can intervene — where an automated system's response triggers another automated response in a spiral no one intended. Managing that risk requires deliberate design: keeping humans meaningfully in control of consequential decisions, building in safeguards and the ability to pause, and resisting the temptation to hand ever more authority to machines simply because they are faster. The wisdom to use AI well may prove harder to build than the AI itself.

The data advantage and its discontents

Artificial intelligence is, at bottom, a technology of data — its systems learn by digesting vast quantities of examples, and their capability is bounded by the quantity and quality of the data they are trained on. This gives India a peculiar mix of advantages and vulnerabilities. On the advantage side, India generates enormous quantities of data across its digital economy and its vast surveillance requirements, and it possesses one of the world's largest pools of the talent needed to turn that data into working systems. On the vulnerability side, military data is scarce, sensitive and hard to share, and building AI for defence means solving the problem of how to train capable systems on the limited, classified, carefully guarded data that military applications involve.^[1]

The challenge is compounded by the gap between the civilian and military worlds. India's AI strength lives overwhelmingly in its commercial technology sector — in start-ups, IT firms and research institutions working on civilian problems — while the military's needs are specialised, secret and structured around procurement processes that move far more slowly than the technology. Bridging this gap, drawing the country's formidable civilian AI talent into contact with defence problems without compromising security, is one of the central institutional challenges of military AI, and much of India's effort — the funding of defence AI start-ups, the challenges and partnerships aimed at industry — is designed precisely to build those bridges.

There is also the matter of trust in the data itself. An AI system is only as good as what it learned from, and an adversary who can corrupt that learning — feeding poisoned data into the training process, or manipulating the inputs a fielded system sees — can turn a military's own algorithms against it. Ensuring that military AI is trained on trustworthy data and is robust against deliberate manipulation is not a peripheral concern but a foundational one, and it is among the harder problems in the whole field.

Decision support and the human commander

Among the most consequential and least understood applications of military AI is decision support — the use of algorithms not to pull a trigger but to help a human commander understand a situation and choose among options. Modern commanders face a flood of information from countless sensors and sources, more than any human mind can absorb and synthesise in the time available. AI can help by fusing that information into a coherent picture, identifying patterns and threats, modelling the likely consequences of different courses of action, and presenting the commander with clear, timely options rather than an overwhelming torrent of raw data.^[1]

Used well, this is a profound enhancement of human command — it lets a commander see more clearly, decide faster and better, and act with a fuller understanding of the situation. But it also carries a subtle danger that thoughtful militaries worry about: the risk that commanders come to over-trust the machine, to accept its recommendations without the scrutiny and judgment that command demands, to let the algorithm's confident output substitute for their own thinking. An AI decision-support system that is confident and wrong — and AI systems can be exactly that, failing in ways that are hard to predict — could lead a commander badly astray precisely because he has learned to rely on it. Preserving genuine human judgment atop the machine's assistance, ensuring the commander remains the decision-maker rather than the rubber stamp, is a design and cultural challenge as important as the technology itself.

India's approach, like that of most serious militaries, emphasises AI as an aid to human decision-making rather than a replacement for it — a tool that informs and accelerates the commander's judgment while leaving the judgment itself firmly human. Holding to that principle as the systems grow more capable and their recommendations more compelling will be a continuing discipline, but it is the right one: the machine should make the commander wiser and faster, not make his decisions for him.

The unglamorous revolution: logistics and maintenance

Much of the public conversation about military AI fixes on its most dramatic and frightening applications — autonomous weapons, killer robots, algorithmic warfare — but some of the most valuable uses are quiet, unglamorous and likely to deliver real benefits far sooner. Chief among them is predictive maintenance: using AI to analyse the streams of data that modern equipment generates and to predict, before it happens, when an engine, a vehicle or a system is likely to fail. For a military operating vast quantities of complex and expensive equipment, often in harsh conditions and remote places, the ability to anticipate failures and fix them before they occur is enormously valuable — it keeps more equipment available, reduces cost, and prevents the catastrophic breakdowns that can cost lives.^[4]

AI-driven logistics offers similar quiet gains. A modern military is, in large part, a vast logistical enterprise — moving fuel, ammunition, spares, food and equipment across enormous distances to where they are needed, when they are needed. Optimising that flow, predicting demand, routing supplies efficiently, and managing the immense complexity of a military supply chain are exactly the kinds of problems AI excels at, and getting them right can be the difference between a force that can sustain itself in a long campaign and one that runs out of what it needs. For a country like India, with forces deployed across mountains, deserts, coasts and high-altitude posts where every delivery is costly and difficult, the logistical payoff of AI may ultimately exceed the value of any autonomous weapon.

These applications matter for a strategic reason beyond their immediate utility: they are where AI can deliver value soonest, with the least ethical peril and the fewest technical risks, building the institutional experience, the data infrastructure and the trust that more advanced applications will eventually require. A military that learns to use AI well for maintenance and logistics is a military laying the foundations — in data, in skills, in culture — for the harder applications to come. The unglamorous revolution is, in this sense, the essential first chapter of the whole story.

Autonomy and the spectrum of human control

The question of autonomy — how much a machine should be allowed to do on its own — is not a simple binary but a spectrum, and understanding that spectrum is essential to grasping both the promise and the peril of military AI. At one end lies full human control, where a person makes every meaningful decision and the machine merely executes. At the other lies full autonomy, where the machine perceives, decides and acts entirely on its own. Between these extremes lies a vast and contested middle ground, and it is in that middle ground that most of the practical and ethical questions of military AI actually reside.^[5]

The prevailing principle in most serious militaries, India's included, is that of meaningful human control — the insistence that a human being remains genuinely in or on the loop of any decision to use lethal force. The machine may sense, analyse, recommend and assist; it may take on the dull and dangerous tasks; but the decision to take a human life remains with a human being who bears the moral and legal responsibility for it. This principle preserves accountability, guards against catastrophic error, and honours the conviction that the choice to kill is one that a machine should never make alone.

Yet holding to meaningful human control grows harder as the technology advances and as the tempo of war accelerates. In a domain where machine speed may decide survival, there is relentless pressure to delegate more to the machine, to shorten the human's role, to let the algorithm act because the human cannot act fast enough. An adversary who abandons human control for the advantage of machine speed poses a genuine dilemma to one who does not. India's challenge — and the world's — is to preserve meaningful human control over the gravest decisions even under the pressure of competition and speed, resisting the temptation to hand ever more authority to machines simply because they are faster. This is less a technical problem than a matter of discipline and values, and it may prove one of the defining moral tests of the age.

The struggle to write the rules

Because the ethical stakes of military AI are so high, there has been a sustained international effort to establish norms, limits and perhaps binding rules — particularly around lethal autonomous weapons — and India's position within that effort reflects its characteristic balancing of principle and interest. For years, states, scientists, humanitarian organisations and advocates have debated in international forums whether and how to constrain weapons that could select and engage human targets without human control, with proposals ranging from outright prohibition to codes of responsible conduct.^[5]

The debate has proven difficult to resolve, for reasons that illuminate the whole problem. The technology is dual-use and evolving rapidly, making it hard to define precisely what should be banned or limited. The major military powers are reluctant to forswear capabilities their rivals might pursue, creating a familiar dynamic in which mutual suspicion frustrates restraint. And the very definition of an autonomous weapon, and of meaningful human control, resists easy agreement. The result has been extensive discussion but limited binding constraint, leaving the development of military AI to proceed largely ahead of the norms that might govern it.

India has generally engaged with these discussions constructively, supporting the responsible development of military AI and the principle of human control while guarding, as ever, its strategic autonomy and its interest in not being constrained while others advance. This balanced posture — advocating responsibility and human control in principle, while ensuring India is not left behind in practice — mirrors its approach across many domains of emerging military technology. Whether the international community can eventually build meaningful norms around military AI, before the technology outruns the possibility of governing it, is one of the great open questions of the field, and India, as a major power with both formidable AI talent and a stake in a stable order, will be an important voice in answering it.

The hidden value: training, simulation and wargaming

Among the applications of military AI that attract little public attention but offer substantial value is its use in training, simulation and wargaming — the preparation of forces and commanders for the conflicts they may one day face. Modern militaries increasingly rely on sophisticated simulations to train their people, rehearse operations, and test plans and strategies, and artificial intelligence can make these simulations far more realistic, responsive and valuable. AI can generate intelligent and adaptive opposing forces that behave like a thinking enemy rather than following a fixed script, forcing trainees to confront genuine unpredictability.^[1]

This matters because the quality of a military is shaped profoundly by the quality of its training, and realistic, challenging training produces forces better prepared for the confusion and difficulty of real operations. AI-enhanced simulations can expose commanders and units to a wide range of scenarios, test their responses, and help them learn from mistakes in a setting where errors cost nothing but teach much. They can model complex situations, explore the likely consequences of different strategies, and help planners understand problems too complex to grasp intuitively. For a large military that must prepare for a wide range of contingencies across demanding and varied environments, the value of better training and planning tools is considerable.

Wargaming — the structured exploration of hypothetical conflicts to test strategies and reveal problems — similarly benefits from AI's ability to model complexity and generate realistic dynamics. Used well, AI-enhanced wargaming can help a military think through the challenges it might face, identify weaknesses in its plans, and prepare for possibilities it might otherwise overlook. These applications carry none of the ethical weight of autonomous weapons and little technical risk, yet they contribute meaningfully to the readiness and effectiveness of a force by improving how it trains, plans and prepares. They exemplify the quieter, less dramatic dimension of military AI — the ways it can make a military better not by fighting for it but by helping its people learn, plan and prepare more effectively, building the human competence on which all military success ultimately rests.

The two-way spillover of a dual-use technology

Artificial intelligence is the quintessential dual-use technology, and the flow of capability between its civilian and military applications runs powerfully in both directions — a dynamic that shapes India's whole approach and offers it a distinctive advantage. The algorithms, techniques and expertise developed for civilian purposes — for commerce, healthcare, transport, science — transfer readily to military applications, and the reverse is equally true. This means that a nation's civilian AI strength is, in large part, latent military strength, and India's formidable civilian technology sector represents a deep reservoir of capability that can be drawn upon for defence.^[2]

This dual-use character distinguishes military AI from most previous military technologies, which were developed in dedicated defence establishments largely separate from the civilian economy. AI, by contrast, is advanced overwhelmingly in the commercial world — in technology companies, start-ups and universities working on civilian problems — and the cutting edge often lies outside the traditional defence sector entirely. The challenge and the opportunity for India both flow from this: the opportunity to harness one of the world's great civilian AI ecosystems for national security, and the challenge of building the bridges that connect that fast-moving commercial world to the slower, more secretive world of defence.

India's efforts — the funding of defence AI start-ups, the challenges and partnerships aimed at drawing industry into defence problems, the institutions created to promote military AI — are largely aimed at building exactly these bridges. The goal is to create the pathways by which India's civilian AI talent and innovation can flow into military capability, and by which the demands of defence can draw on the strengths of the commercial sector, without compromising security. Getting this right — connecting the two worlds effectively — is one of the central challenges of India's military AI effort, and success would allow the country to translate its genuine and considerable civilian AI strength into a corresponding military advantage. In a domain defined by dual-use technology and civilian-led innovation, the nation that best bridges the civilian and military worlds may hold the decisive edge, and India's deep civilian AI ecosystem gives it a foundation from which to build one.

The imperative and the caution

India stands, then, at a characteristic crossroads. The strategic imperative is undeniable: in a world where its rivals are racing to build military AI, India cannot afford to fall behind, and its own formidable technological talent gives it every reason to compete. The country is building the institutions, funding the research, drawing in its start-ups, and applying AI where it most helps — in seeing the battlefield, in enabling autonomy, in defending its networks, in keeping its machines running.

Yet the same technology demands caution — about the ethics of autonomous force, about the brittleness of algorithms, about the dangers of machine-speed escalation. The mature course, and the one India appears to be steering, is to pursue military AI vigorously while insisting on human control of the gravest decisions and on the robustness and security of the systems it fields. The thinking battlefield is coming whether any nation wills it or not. The question for India is whether it can be both fast enough to keep pace and wise enough to keep control — and on that balance may rest not only its security but its conscience.