India's AI Mission: Baby steps
China is ahead of the US in 57 of 64 critical technologies that shape strategic power
India has a long way to go in AI
Without fostering quality education across the board, so as to produce high-quality research in critical technologies, India cannot protect strategic autonomy
If wishes were horses, India’s Artificial Intelligence Mission would catapult India into the world’s major AI league, right there with the US and China, with India’s own foundation model developed in just 10 months. India is going to make available cloud computing capacity to researchers and industry at ultra-low costs per compute (jargon for a single computational operation, like running a specific algorithm). The idea is sound, and needs to be implemented, and scaled up. But it will not prevent India still being a laggard in AI.
The government will ask a clutch of providers of powerful cloud computing capacity, represented by advanced graphic processing units (GPUs), the moniker for chips on which AI runs, to share their infrastructure to make computing power available for users through a common portal. The government will subsidise the cost of this infrastructure. The proposal at present is that a bunch of Sarkari babus will vet the applications for access to this capacity. And a chosen set of startups would use this compute infrastructure to develop assorted AI models and applications.
After all, China’s DeepSeek has demonstrated that high-quality AI can be developed at a fraction of the funds leading American companies have been burning through for the task. Why should India lag, if China can succeed?
The Australian Strategic Policy Institute offers a clue. This body has accomplished the remarkable task of identifying 64 (it started off with 44) critical technologies and tracked the progress of different nations in generating research in these areas of strategic capacity. They have a project, and a web site, called Critical Technology Tracker – the global race for future power.
These technologies belong to nine broad clusters: AI technologies, Advanced information & communication technologies, Advanced materials & manufacturing, Biotechnology, gene technologies & vaccines, Defence, space, robotics & transportation, Energy & environment, Quantum, Sensing, timing & navigation, and Unique AUKUS technologies.
AUKUS is a sub-group of the Five Eyes Alliance, comprising Australia, the United Kingdom and the United States (the other two sets of eyes belong to Canada and New Zealand). By unique AUKUS technologies, the Institute means Air Independent Propulsion, Autonomous Underwater Vehicles, and Electronic Warfare. Why these are designated as unique AUKUS technologies is not obvious, probably, the AUKUS allies had a special project to develop these technologies.
ASPI then identifies the key technologies in each cluster, and measures the contribution of different nations to developing expertise in each area.
Below, we present ASPI’s findings, just for the AI cluster. This gives you a comparison of research quantity and quality. The percentage of publications in total and the percentage of publications in the top 10% of highly cited papers in this technology are shown as indicative metrics. The figure in parentheses reflects whether or not the contribution of quality research is greater or smaller than the contribution of quantity and by how much—when the quality proportion is higher than the quantity proportion, the figure is positive, and when the quality proportion is lower than the quantity proportion, the figure is negative.
In all these areas, China’s total research output is several multiples of India’s. The gap is wider when it comes to quality publications. Chinese quality is high, while the quality of Indian papers is, on the whole, low, as indicated by the negative quantity-quality gap for India and the absence of a positive gap even in any one area.
China leads in the other technology clusters as well, and by a wide margin.
If it is any consolation, it is not just India that China beats in research output. ASPI concluded, in a report dated August 30, 2024 that China is ahead of the US in most critical technologies: “China led in just three of 64 technologies in the years from 2003 to 2007, but is the leading country in 57 of 64 technologies over the past five years from 2019 to 2023. This is an increase from last year’s Tech Tracker results, in which it was leading in 52 technologies.”
The short point is that capability and leadership in critical technologies that determine strategic dominance is achieved not by offering some state-subsidised compute facility for a select set of players, chosen by all-knowing bureaucrats, but by fostering quality education from kindergarten to doctoral research, in diverse areas.
In India, policymakers are in a hurry to reduce universities to imparters of ‘skills’, instead of treating them as hallowed sites where the frontiers of knowledge are constantly pushed outward, for the collective benefit of humanity, while also yielding specific spin-offs to national advantage.
The AI initiative is welcome. But learning to kick a ball straight is just the beginning, not a feat that lets you compare yourself to Lionel Messi.
Our approach to science is flawed on many counts. The religious overtone that suppressess genuine research is a major one that pollutes the ecosystem. For the best brains that come out of our IITs, there is hardly any worthwhile occupation or opportunity in the country. So, we end up as suppliers of the best professionals to the US and others. And we pay a heavy price for importing the technologies and services developed by them!