In August 2010, after years of interrogation of terrorists and terror suspects, and through intelligence reports, surveillance, wiretapping, and satellite images and videos, the Central Intelligence Agency (CIA) finally tracked the al-Qaeda head to an Abbottabad compound in Pakistan’s Khyber Pakhtunkhwa.

On May 2, 2011, the Joint Special Operations Command’s SEAL Team Six eliminated the al-Qaeda head.

AI’s Use In Tracking Osama      

It was humanly impossible to structure and analyse the treasure trove of data collected over the years and establish a pattern to locate bin Laden.

Palantir Gotham, the defence and intelligence software of US company Palantir Technologies, did the job. 

Gotham used artificial intelligence (AI) to integrate, process, and analyse the data and identify patterns and connections. After connecting the dots, the CIA concluded that bin Laden was in the compound. 

According to Palantir, “Gotham’s targeting offering supports soldiers with an AI-powered kill chain, seamlessly and responsibly integrating target identification and target effector pairing.”

Palantir Gotham’s clients include the CIA, National Security Agency (NSA), Federal Bureau of Investigation (FBI), Marine Corps, the US Air Force, Special Operations Command, Pentagon, and Immigration and Customs Enforcement, among others. 

Sixth Domain Of Warfare  

The use of AI for military purposes can be broadly divided into two categories.

Firstly, Lethal Autonomous Weapon Systems (LAWS) such as drones and robots. Automated drones aren’t operated remotely, but act independently once activated.

US defence technology company Anduril Industries makes autonomous unmanned aerial systems (UAS), counter-UAS (C-UAS) and surveillance systems. 

Secondly, Decision Support Systems, software apps that analyse intelligence, plan, and manage battles, communications and logistics. 

Anduril’s software platform, Lattice, is the US Army’s fire control platform for C-UAS missions. It gathers input from several sensors and systems and integrates them. Subsequently, AI, machine learning algorithms, and data processing analyse the incoming information. 

Palantir’s Maven Smart System integrates disparate surveillance and intelligence data from drones, satellites and other sensors into a single interface to speed up military targeting. 

Similarly, American AI company Anthropic’s Claude family of large language models (LLMs) assesses intelligence, prioritises targets, and simulates battle scenarios after processing huge volumes of drone, satellite, and signal data.

With the Maven-Claude integration, the US had the most lethal AI weapon. While Claude reasons and summarises, Maven identifies the target. 

The US used the integration to capture Venezuelan President Nicolas Maduro, and even bomb 1,000 targets. This includes the assassination of Iran’s Supreme Leader Ayatollah Ali Khamenei in the first 24 hours of the war.

Other AI Decision Support Systems are Israel’s Lavender and Gospel, which were used in the Gaza War.

AI In Operation Sindoor

The Indian Army achieved a milestone during Operation Sindoor by using AI to cut the kill chain and achieve 94% accuracy. 

The Army used 23 apps, including the Electronic Intelligence Collation and Analysis System (ECAS) and TRINETRA, integrated with Project SANJAY, according to Lieutenant-General Rajiv Kumar Sahni, Director General (DG) of Electronics and Mechanical Engineers.

ECAS, which has a 26-year-old database of frequency signatures and multi-source intelligence, identifies, analyses, and prioritises the enemy’s radars and surveillance sensors to enhance the Army’s decision-making. 

The TRINETRA System is a battlefield surveillance and intelligence platform that creates AI-powered heat maps and combines multi-source data to identify enemy sensors and track targets.

SANJAY, an automated system, integrates inputs from ground and aerial sensors, processes them, and finally fuses them to produce a common surveillance picture of the battlefield.

The integration of TRINETRA with SANJAY created a common operational and intelligence picture, enabling commanders to improve their situational awareness and make faster decisions.

India Needs To Catch Up Fast

India has made noticeable progress in AI.

According to Stanford University’s Artificial Intelligence Index Report 2025, India was ranked third in the Global AI Vibrancy rankings, with a score of 21.59 in 2024, ahead of South Korea (17.24) and the UK (16.64). 

The US was ranked first with a score of 78.60, and China was in second spot at 36.95.

In research and development (R&D), India was third with a score of 7.13. The US was first (19.15), followed by China (17.22).

However, in global private investment in AI in 2013-24, India ranked seventh at US$11.29 billion, the US first (US$470.92 billion), and China second (US$119.32 billion). 

In 2024, India ranked 12th with only US$1.16 billion, and the US first (US$109.08 billion).

In AI’s use on the battlefield, the US and China lead because they tap private tech companies. 

Maven was born out of Project Maven, a US Department of Defence initiative launched in April 2017. Private tech companies, from Google and Palantir to Anduril and Anthropic, have supported Project Maven. 

China aims to turn AI into a US$100 billion industry by 2030. 

Under Xi Jinping’s “civil-military fusion”, private AI companies are backed by the state, including a US$8.2 billion AI fund for start-ups. 

China is catching up to the US very fast, according to the Stanford University report.

In 2024, US-based institutions produced 40 notable AI models, outpacing China’s 15 and Europe’s 3. However, “Chinese models have rapidly closed the quality gap”.

Performance differences on Massive Multitask Language Understanding (a prominent AI benchmark designed to evaluate LLMs) and HumanEval (an OpenAI-developed benchmark dataset for evaluating LLMs) shrank from double digits in 2023 to near parity in 2024 between the US and China. 

Battle Beyond Autonomous Weapons 

India’s progress in LAWS (AI autonomous weapons) is noteworthy. 

The Army uses the Ten AI Weapon System at the LoC to stop infiltration and the Multi-Utility Legged Equipment. DRDO is developing a humanoid robot that will be on the frontline in high-risk situations to reduce troop exposure.

Under the Chanakya project, multiple UAVs operate together. The Kaala Bhairava E2A2 is the country’s first AI-powered medium-altitude long-endurance autonomous combat drone. Nagastra-1 is India’s first indigenous kamikaze drone. Sheshnaag-150/Sheshnaag-156 is another kamikaze drone. 

However, the real competition in using AI on the battlefield is the integration of LLMs with military systems or military-specific LLMs.

China, which already excels in autonomous drones, namely Hongdu GJ-11, Jiu Tian and AR-2000 Helicopter Drone, is integrating advanced LLMs with its military architecture.

China’s Xian Technological University used the famous Chinese generative AI chatbot DeepSeek to generate 10,000 military scenarios in only 48 seconds, the South China Morning Post reported last May.  

In 2024, PLA-linked Chinese research institutions incorporated their own parameters into Meta AI’s open-source Llama 13B LLM to make ChatBIT, a military-specific AI tool that gathers and processes intelligence and assists in operational decision-making, according to Reuters, which read the research paper. 

The PLA’s “War Skull” is a wargaming system that adapts to different kinds of enemies. 

The use of AI in the Iran War is just the beginning. The next phase will be “intelligentised warfare”: a war between algorithms. 

(The writer is a columnist with more than two decades of experience in journalism. Views expressed are personal.)