India's Artificial Intelligence Summer: Is it even taking off? The Good, the Unfortunate and the Uncertain
From Abhivardhan, our Chairperson
This is a post authored by Abhivardhan, our Founder and Chairperson.
The insight by Amit Sethi, Prof. at Indian Institute of Technology, Bombay is gold.
He says that the best of the best AI talent as per his understanding - in India - are still stuck with fine-tuning existing AI models, and not building cutting-edge AI in India.
I agree with him. The US Innovation Ecosystem is ahead of India here for obvious reasons that the Indian system is still not equipped and things may take time.
Here are some points on why this is a problem for India's AI dreams.
The Indian AI "summer" (also known as India's GenAI Funding Summer), which was reported by Analytics India Magazine (read: https://www.linkedin.com/pulse/indias-genai-funding-summer-begins-analytics-india-magazine-ams4c/?trk=public_post_main-feed-card_feed-article-content) is yet to mature to produce Artificial Intelligence use cases that credible.
I completely agree with the practical vision of Sh. Nandan Nilekani, the maverick behind India's Aadhar - that India must not emulate to create big models like OpenAI, which has its own deleterious issues and problems - but must develop micro-level or smaller yet effective GenAI use cases (read insight by moneycontrol.com: https://www.moneycontrol.com/news/technology/indias-cto-nilekani-unveils-ai-strategy-focus-on-use-cases-not-my-model-bigger-than-yours-11853961.html). The problem is we do not know what kinds of AI model weights can be justified as smaller AI use cases which have limited application, can be standardised and work for long.
India's tech policy on AI - including those on IndiaAI cannot succeed unless we encourage India's local capabilities have to be prioritised. We are yet to see if India's tech ecosystem is focusing on those aspects. The Government is also yet to make sense on this. In fact, an American semiconductor company CEO also stated "if India has to capitalize on this revolution, then it has to be Indian homegrown companies. MNCs may setup engineering centres and provide jobs, but that doesn't create wealth or IP. Homegrown companies do."
The current GenAI use cases developed by some major Indian companies are sadly getting into the stage of becoming knock-offs or heavily reliant upon the models built by OpenAI, Anthropic and others, which should be seriously avoided.
Thanks for attending my TED talk. People usually share a lot of stuff on India's AI landscape. I hope my contrarian yet unique points might add something fruitful to your discourse.
You can also know more about India's AI and Law situation by reading my book: