The world is going gung-ho over the Generative AI models as industry verticals are latching on to the LLMs (Large Language Models) to increase productivity by tapping the humongous amounts of data within and outside the organisations.
Is India lagging behind in this space?
While the US and China are making strides in research and deploying Generative AI models, India, by far, remains a bystander except for the small projects that are happening here and there.
An LLM model can give human-like answers, providing contextual references based on the data that they are fed on.
“While the research and development in this space is led by private corporations in the US, Government-led initiatives are pushing the Generative AI research in China. There is a feeling that India is left behind,” Jaspreet Bindra, Founder-Managing Director of the UK-based Tech Whishperer, told Businessline.
“We think there is an immediate need to promote a Made-In-India GenAI or JanAI, as we call it. LLM brings in the Indian context by tapping into the data,” he said.
“The land is called in different ways in different parts of the country. Over years, we built humongous amounts of data which is largely not factored in global Generative AI models. We need something that captures the local context,” he said.
Bindra, along with Sudhir Tiwari, Global Head of the Digital Engineering Centre at Thoughtworks India, mooted the idea of ‘JanAI’, providing a basic framework for the proposed Generative AI model for India.
India, they argued, can follow a different model than that of the US and China. “We have already demonstrated our ability to build something like Aadhar that proved that we can build something at a national scale.
“An exclusive JanAI or GenAI for India can be developed as a digital public good,” Tiwari said.
He, however, felt that the quality of data was very important as the country builds a foundational level,he said.
Asked about the cost of setting up such a huge model requiring computational infrastructure, they said the cost of computing was coming down and when a public-private collaboration is formed, it should remove the problems.
