26 January 2024

Early steps in India’s use of AI for defence

Antoine Levesques

India is just starting to use artificial intelligence for national-defence purposes, but it plans to increase these capabilities by working with domestic industry and through burgeoning overseas partnerships.

In November 2023, at a virtual G20 summit hosted by India, Prime Minister Narendra Modi spoke about global concerns regarding the ‘negative use’ of artificial intelligence (AI). These remarks were slightly more pessimistic than those made days earlier by his information-technology minister, Rajeev Chandrasekhar, at the Bletchley Park AI Safety Summit in the United Kingdom. Despite any general concerns the government may have regarding AI safety, it is pushing ahead to find military uses for the technology to remain globally competitive, notably as part of the sharpening strategic contest in the Indian Ocean region between New Delhi and Beijing. China has a globally competitive AI sector and could field new military applications of this technology using a relatively small percentage of its defence budget (perhaps as little as 1–2% in the coming years). Pakistan, meanwhile, is also taking steps to acquire AI capabilities for national defence.

The contest to develop national AI capabilities has sharpened significantly since the public launch in late 2022 of highly capable large-language models. India’s current efforts, however, are building on work that began long before this proof of concept. India launched its first national strategy for AI in June 2018 at roughly the same time that an AI task force convened by the government delivered defence-specific recommendations. Those included the creation, effective in 2019, of a high-level Defence AI Council and a Defence AI Project Agency.

In 2022, the government published a list of 75 priority projects related to using AI for defence; these focused on data processing and analysis, cyber security, simulation and autonomous systems, particularly drones. India is also exploring AI applications for underwater domain awareness and border security. And ongoing work to incorporate AI solutions into its civilian space programme could have indirect defence applications. Some officials have privately dismissed as unlikely the idea that India will seek to use AI for decision-making related to the lethal use of force. Amit Shah, India’s interior minister, has highlighted the key role of AI in counter-terrorism investigations and prosecutions. Meanwhile, at an AI summit in New Delhi in December 2023, Prime Minister Modi emphasised the need to prevent the misuse of AI technology by terrorist organisations. In the near term, India is prioritising the needs of its domestic industry to help it become globally competitive in the AI sector. The government hopes to harness the private sector’s large AI workforce and dual-use products both in defence and in its plans to double (from 11% to 22%) the digital economy’s contribution to GDP by 2026.

Early defence efforts 

In 2021, the Indian Army demonstrated an AI-enabled swarm of 75 aerial drones and used AI for intelligence, surveillance and reconnaissance purposes in that year’s Dakshin Shakti military exercise. While the number, nature and results of public–private programmes related to AI are classified, the government is already running projects related to predicting atmospheric visibility, imagery analysis, drone-collision avoidance and ship tracking. AI-application centres embedded in each of the three armed-service branches – at the Military College of Telecommunication Engineering, Mhow (Army), the INS Valsura (Navy) and Air Force Station Rajokri (Air Force) – are supporting many of these efforts.

India’s long-standing public-sector defence organisations are playing a leading role regarding AI; these include Bharat Electronics Limited and the Defence Research and Development Organisation (DRDO). These organisations are adept at working with the military branches and civilian higher-education institutions. Public-sector research and development (R&D) bodies in India often fail to deliver programmes on time and to service specifications (although they completed 26 out of 61 projects in 2022). Reforms likely to move forward at the DRDO, such as creating new links to academia and small businesses, could prove beneficial to AI research. However, their progress will be hampered by the recent downward trend in R&D allocations in the defence budget, which has been met with calls to be reversed. Overall, R&D as a share of GDP is low in India.

International partnerships 

India is including AI in major ongoing and next-generation defence partnerships to support its growing domestic industry. This also takes advantage of the lack of international trade rules for AI products related to national security.

United States Secretary of Defense Lloyd Austin and his Indian counterpart first met to discuss US–India cooperation on defence AI in November 2022. US President Joe Biden and Modi had announced bilateral AI cooperation through the US–India Initiative on Critical and Emerging Technologies in May 2022, which has since been launched at the national-security-adviser level. In May 2023, the two countries held their first dialogue on advanced domains, and the following month, they agreed to collaborate more closely on high-performance computing.

India’s DRDO also signed an agreement with its Israeli counterpart in 2021 to develop technologies including AI, and it may be considering similar cooperation with France, Italy and the UK. Meanwhile, other countries, including Russia, are already positioning AI-enabled products of their own to win a greater share of the sales and co-creation opportunities in India. There is no evidence, however, that India has partnered with particular countries to acquire specific defence AI capabilities, as it does in other defence trade and industrial partnerships. It is also unclear how much India’s broad ‘Make AI in India and Make AI work for India’ vision relies on cooperation agreements established as part of its ‘Make in India, for the world’ manufacturing initiative, which it increasingly prioritises in its next-generation overseas defence-industrial partnerships.

On the diplomatic stage, India has said that it supports openness, safety, trust and accountability on AI, echoing the rhetoric of many of its traditional partners, including Russia, but also China. Although the military is aware of both the opportunities and threats of defence AI, India does not yet have a clear view of how it could affect strategic stability or whether its nuclear doctrine should change as a result. This is significant given tensions with its nuclear-armed neighbours, China and Pakistan. Ongoing work to publicly release India’s first national-security strategy – possibly after the April-May 2024 general election – could offer more clarity and set expectations in the longer term for an AI strategy for defence.

No comments: