Infosys Co-Founder Nandan Nilekani Applauds China's DeepSeek for Open AI Innovation

By siliconindia  |  Monday, 10 March 2025, 10:28 IST

Infosys Co-Founder Nandan Nilekani Applauds China's DeepSeek for Open AI Innovation
Infosys Co-Founder Nandan Nilekani has denounced the practice of creating more sophisticated artificial intelligence (AI) models, saying that this does not always result in improved outcomes. Addressing the People+AI Mela in Bengaluru, he questioned how many AI models are closed systems that the public has no access to.
"They just continue to do more of the same, but it is not that things are going to improve. They are all generally contained within the four walls of a company. It is a black box for all of us," Nilekani said, quoting The Economic Times.
This opaqueness, he cautioned, is constraining innovation and withholding wider contributions to AI development. While criticizing the secrecy of much of AI model development, Nilekani applauded China's handling of AI development, especially that of the DeepSeek project. He noted that China has succeeded in developing powerful AI models using fewer resources but being transparent regarding their approach.
“They have done the world a service. So that goes back to commoditisation”, he stated.
According to Nilekani, AI models are increasingly becoming commodities, with hundreds of new ones emerging worldwide. The openness demonstrated by China’s DeepSeek project, which publicly shared its AI-building practices, has contributed to demystifying AI development.
"I always thought models were a commodity. And that's coming to be true. Hundreds of models are being built. China's DeepSeek demystified model-building. Last week they posted all their open-source practices on Twitter. The knowledge to construct models is being commoditised at a very fast pace", he added.
Outside model-building, Nilekani prioritized high-quality datasets, particularly in Indian languages, to assist AI development in India. He mentioned AI4Bharat, an effort by IIT Madras aiming at open-source AI development in Indian languages, as a stepping stone in that direction.
"The difficult thing is uses. How do we make use of this to bring value to people? That is where we can set the direction. We can always substitute the model", he explained.
AI4Bharat has been supported by Nilekani for the past few years and will remain supported by him for another three years at least. The initiative is focused on increasing AI availability for various linguistic communities in India, filling a significant void in AI research.
Nilekani also spoke briefly of the breakneck pace at which India's startup ecosystem had expanded. There has been growth from 1,000 companies in 2016 to over 150,000 in more recent times fueled by the adoption of digital technologies and rising inflows of money.
"They are going to take mass-scale technology that we've applied in unique use cases. There is going to be tons of capital going in as well", he noted.
In spite of AI models evolving, Nilekani opined the actual challenge comes in their implementation. The emphasis, he maintained, had to be more on how one could use AI to solve actual problems and enhance lives rather than merely constructing sophisticated models.

ON THE DECK