Artificial Intelligence shaping the future of military warfare

Lt Col Akshat Upadhyay

The use of artificial intelligence (AI), which relies heavily on data, algorithms and computation, was seen in the ongoing Russia-Ukraine conflict. AI as a field started as an attempt to create machines that could think or mimic human intelligence. It has diversified and been through phases since its start in the 1950s. Initial attempts focused at identifying different facets of human intelligence — planning, reasoning, problem solving, perception and natural language understanding, solving them separately and finally fitting them together. The lack of advanced computing power and data, hindered it. In the 1980s, systems were created with embedded domain knowledge.

A third attempt at AI — the one operating today — uses data to help machines identify patterns, learn, perform specific tasks and predict. This is known as machine learning (ML) and requires substantial human labour in terms of classifying, cleaning and labelling data to be fed to algorithms. One of its subsets is deep learning (DL) which uses a neural network or a simulation of how human brain seemingly functions to find patterns in unstructured data. However, DL needs a massive amount of data to find patterns.

Today, AI is synonymous with facial recognition, robotics, lethal autonomous weapons systems (LAWS), pattern recognition, data fusion and predictive maintenance in the military realm. Underlying these advances is algorithms of ML and DL. Facial recognition is a force multiplier for troops in counterinsurgency and counterterrorism operations. AI can help segregate terrorists from ordinary citizens. Combined with other indicators like last known geolocation, call history and human intelligence, this data can create a “pattern of life” profile. The US used a similar system in Afghanistan to monitor and, if needed, liquidate high-value targets.

Ukraine contracted Clearview, a US firm specialising in facial recognition for detecting infiltrator checkpoints, identifying more than 2,30,000 Russian soldiers and officials who have taken part in the conflict, prosecute members of pro-Russia militias and collecting evidence of Russian war crimes.

Extending the autonomy of weapons systems is another AI usage, especially unmanned aerial systems (UAS). Autonomous systems without AI are used such as loitering munitions (LMs), the presence of AI “provides” aerial systems with a controversial of capability — to select and target humans without any intervention from humans. Emphasis is on “provide” since most AI-based autonomous systems use their autonomy only to select a target, the decision to kill still rests with a human. However, in Ukraine the Saker Scout drone was reportedly used to kill Russian soldiers in an autonomous mode.

The Indian Army has started of absorbing AI within the organisation. Close to 140 AI-based surveillance systems have reportedly been deployed along the disputed borders. Real-time natural language processing (NLP) systems, translating Mandarin to Hindi, English and local languages are being field-tested.

With increasing use of AI and autonomous systems by India’s adversaries, forces need to transition to embrace this technology, own it to transform structures and doctrines.

A recent iteration of AI in the commercial realm is generative AI which uses generative adversarial networks to create new images and videos. Combined with the transformer model based large language models, this has given rise to a new category of virtual reality (VR) where users can dynamically interact with new characters and which can respond to the user’s questions and actions. VR and AI, when combined together for the military can be used to simulate combat scenarios – stress-testing soldiers for real-life operations. VR is also used for understanding, learning and operating on many platforms and inculcating new skills. Commanders are learning and making operational plans in VR. In effect, VR as a training, learning and planning aid holds great potential for the military.

The writer is a research fellow, strategic technologies, Manohar Parrikar Institute for Defence Studies & Analyses