India is hosting a high-profile AI Impact Summit next week — the fourth in a series of global summits. When the first of these launched in Bletchley Park in the UK in 2023, the focus was all on AI safety; today, many of those worries have been tossed out as we witness a US-China race, with the EU vying to play referee. With this being New Delhi’s turn, the theme is sarvajana hitaya, sarvajana sukhaya — welfare for all, happiness for all.
To be clear, sarvajana sukhaya is a tall order. AI is not ushering in much sukha these days. The early delights of playing with ChatGPT are over. Investors are fretting that the race for the most powerful models will trigger a stock-market crash like no other. College graduates and workers are wondering which jobs will disappear. The multitude of “godfathers” of AI worry about human extinction. But with sarvajana hitaya, India has a genuine shot at leading the world — in AI designed for purpose.
Ironically, while companies hunt for “use cases”, there is no shortage of potential. Consider just three areas: Food, health and human capital. Seventy per cent of the world’s food is produced by low-productivity smallholder farmers; 4.5 billion people aren’t covered by essential healthcare; 739 million adults and 250 million children lack basic literacy. AI can help with each of these.
There are many examples in agriculture alone. The national Kisan E-Mitra chatbot can handle up to 20,000 queries daily in 11 regional languages. Telangana’s AI-led Saagu Baagu project doubled the earnings of chilli farmers by simultaneously boosting yield, cutting pesticide and fertiliser use, and increasing sale prices.
Healthcare offers another opportunity as India’s doctor-to-patient ratio in public hospitals is a horrifying 1:11,000. By mid-2025, the eSanjeevani telemedicine platform had conducted 389 million virtual consultations. Specialised AI platforms such as Qure.ai’s TB detection algorithms have reached millions.
Building skills for a workforce of 950 million of whom only 5 per cent have formal skills training is hard. AI-aided learning platforms can help. Programmes have racked up impressive numbers: 1.6 million on FutureSkills PRIME, with 41 per cent women and 85 per cent from Tier II and Tier III cities; 275 million on DIKSHA with 70 per cent rural penetration.
All that said, the AI Summit should avoid being co-opted into a platform for chest-thumping. India’s leadership will be demonstrated not just by engineering a Ferrari but by rebuilding the rutted road underneath it. We must build the Ferrari while fixing the dirt road. Consider 10 potholes:
Close the internet connectivity gap: Only 24 per cent of rural households have internet access vs 66 per cent in urban areas. Of the 125 countries we study at my Digital Planet research centre, India is close to the bottom in digital gender parity.
Harmonise energy, climate and AI goals: The inadequate energy infrastructure, grid unreliability and weak transmission capacity will be stressed further by the growing demands from AI use.
Make the workforce work-ready: For every 10 AI roles in India, there is only one qualified engineer. Talent is a constraint.
Cut the dependence on foreign supply chains: India imports over 90 per cent of its semiconductor chips along with high-purity chemicals, gases, and silicon wafers. The US-China rivalry has already fragmented the global tech ecosystem. Shifting to hardware manufacturing and compute infrastructure won’t be easy.
Make the data more usable: India generates vast amounts of data, but lacks high-quality, well-annotated datasets, particularly for regional languages.
Streamline AI governance: Regulatory uncertainty with complex customs clearance and documentation requirements that vary across states slow innovation.
Engineer “good enough” AI: AI-for-purpose designs need to be versatile.
Secure the infrastructure: Indian AI can accomplish its scale and reach through DPI that can be extended to AI-powered services. This must be secured.
Bridge the capital chasm for late-stage start-ups: Seed funding for AI startups exceeds Series A through C investments, when companies need resources to scale. While early-stage funding has improved, start-ups still rely heavily on foreign capital beyond Series B.
Fix the actual dirt road: Inadequate infrastructure, traffic congestion, and fragmented operations add costs.
Beyond the US-China-EU trinity, India demonstrates a fourth possibility. We must speak not in terms of AI-for-power measures, but in terms of the yield of the average smallholder farm, the life expectancy of the poorest, the percentage of those who can read this page. That’s the AI race that India can — and ought to — win.
The writer is dean, Global Business, The Fletcher School at Tufts University
