Nandan Nilekani, the co-founder and chairman of Infosys, has expressed concerns about the growing trend of developing increasingly complex artificial intelligence (AI) models, arguing that this does not necessarily lead to improved outcomes. During his address at the People+AI Mela in Bengaluru, he noted that many companies keep their models confidential, resulting in what he describes as “black boxes” for the public. This veil of secrecy, he argues, is hindering progress, as reported by the Economic Times.
Nandan Nilekani commends DeepSeek initiative
Nilekani voiced his admiration for China’s strategy in AI model development. He observed that China has successfully built effective AI systems with fewer resources while maintaining transparency about their processes. He believes such openness facilitates the dissemination of knowledge regarding AI and fosters global advancement. He further commented that AI models are evolving into commodities, with numerous new models continuously emerging. Nilekani highlighted China’s DeepSeek project, which has shared its AI development techniques publicly, thereby simplifying the learning process for others and encouraging innovation. “They have provided a valuable service to the world. This ties back to commoditization,” he remarked.
“I have consistently maintained that models are becoming commoditized. This is increasingly evident. Numerous models are being developed. China’s DeepSeek has demystified the process of model construction. Recently, they publicly shared all their open-source practices on social media. The expertise required to create models is rapidly becoming accessible,” Nilekani added.
He also emphasized the importance of high-quality data, particularly in Indian languages, for enhancing AI development in India. He mentioned AI4Bharat, an initiative by IIT Madras that aims to promote open-source AI solutions for Indian languages. Nilekani has supported this initiative for several years and intends to contribute for the next three years.
In conclusion, while AI models are expected to continue evolving, the primary focus should remain on leveraging AI in practical applications that positively impact the lives of individuals.
