World Bank says India should scale small AI applications instead of betting heavily on frontier models
The World Bank suggests India focus on scaling existing AI applications in healthcare and education rather than heavy investment in frontier models.
The World Bank has proposed a strategic shift for India's artificial intelligence journey, suggesting that the nation should prioritize scaling small, specialized AI applications across the economy rather than pouring massive resources into developing advanced frontier models from scratch.
In an exclusive conversation with ANI, Vincent Belinga, a Senior Economist at the World Bank, outlined a pragmatic roadmap for India to maximize the benefits of artificial intelligence. He argued that while developing cutting-edge technology is important, the immediate priority must be deepening the adoption of existing AI tools to ensure productivity gains are felt by millions of workers and businesses across the country.
The Three-Stage AI Roadmap
Belinga suggested that India's path to becoming an AI powerhouse should follow a structured three-stage approach:
- Adoption: Utilizing existing AI applications to improve efficiency.
- Adaptation: Taking open-source AI models and tailoring them to suit local Indian languages, conditions, and specific needs.
- Frontier Development: Developing the next generation of advanced AI capabilities from the ground up.
The economist emphasized that while adoption is currently on the rise, it is not yet broad-based enough to transform the entire economy. The goal should be to move beyond isolated success stories and ensure that millions of firms and workers can use AI to boost their productivity.
Scaling Local Success Stories
India has already demonstrated significant potential through various localized AI initiatives. However, many of these remain confined to specific states or are stuck in the pilot project phase. The World Bank highlights several promising areas where AI is already making an impact:
- Multilingual Access: Tools like Bhashini and AI4Bharat are breaking language barriers for digital services.
- Education: The DIKSHA platform is providing personalized learning experiences.
- Agriculture: Chatbots like the PM-KISAN tool and KATHIR are assisting farmers.
- Healthcare: AI-driven tools are being used for critical tasks, such as screening for tuberculosis via chest X-rays.
The recommendation is to scale these specific, high-impact applications so they can reach the broader population across all states, rather than leaving them as regional successes.
The High Cost of Frontier Models
While the development of "frontier" AI—the most advanced, large-scale models—is a global race, Belinga warned that it requires astronomical investment. He pointed to the United States, where billions of dollars are being funneled into frontier development.
"Frontier development is very resourceful," Belinga noted, adding that India faces competing development needs. He suggested that India should be highly selective, investing in frontier capabilities only where the country holds a distinct comparative advantage, rather than attempting to compete in every high-cost arena.
Boosting National Productivity
The economic stakes for India are significant. The World Bank estimates that approximately 14 per cent of Indian workers are in roles where AI can complement their work and enhance productivity. Furthermore, depending on how quickly AI is deployed and supported by complementary investments, it could add between 0.1 and 0.8 percentage points to India's annual labour productivity growth over the next decade.
To achieve this, the focus must remain on broad-based deployment. By spreading AI applications across sectors like finance, IT, agriculture, healthcare, and education, India can ensure that productivity gains are distributed throughout the entire economy, rather than being concentrated in just a few high-tech industries.