Nandan Nilekani on AI: Why China’s DeepSeek is a Game Changer

Is AI Innovation Slowing Down Due to Secrecy?

Artificial Intelligence (AI) is evolving rapidly, but many tech giants keep their AI models closed-source—limiting access to their data, algorithms, and breakthroughs. This lack of transparency has raised concerns about slowed innovation and accessibility.

Infosys co-founder Nandan Nilekani has a bold take on this issue. He believes AI models are no longer groundbreaking inventions but commodities, and that open-source AI is the key to progress. One such initiative making waves is China’s DeepSeek, an open-source AI project that challenges the status quo.

Let’s dive into why DeepSeek is a game changer and what India can learn from it.

The Problem with AI Secrecy

Many major AI companies—such as OpenAI and Google—have adopted closed-source policies, keeping their AI models behind paywalls or restricted access. While this helps them maintain a competitive edge, it also:

✔ Slows down AI innovation by restricting knowledge-sharing.
✔ Limits accessibility, making AI development expensive and exclusive.
✔ Creates AI black boxes, where decisions made by AI are difficult to understand and scrutinize.

This secrecy hampers global progress, making AI advancements concentrated in the hands of a few corporations. But DeepSeek is changing this narrative.

What is DeepSeek? Why is it Different?

DeepSeek is China’s answer to open-source AI. Unlike traditional AI models, DeepSeek makes its methodologies, data, and code public, encouraging global collaboration. Here’s why this is revolutionary:

✔ Promotes transparency by allowing experts worldwide to improve and audit AI models.
✔ Accelerates innovation as researchers and developers can build upon existing models instead of starting from scratch.
✔ Demystifies AI development, making it accessible to smaller players, startups, and academic institutions.

With this approach, DeepSeek is setting a new benchmark for AI development—one that favors progress over exclusivity.

Nandan Nilekani’s View: AI Models are Commodities

Nandan Nilekani believes that AI models are no longer exclusive breakthroughs—they are commodities that are being built in hundreds of versions worldwide. With open-source initiatives like DeepSeek, AI is becoming democratized.

His key arguments:

✔ AI models are not special inventions anymore. With the right data and computing power, any company can build them.
✔ The real value lies in how AI is applied—not just in creating bigger models.
✔ Open-source AI projects are key to making AI accessible to developing nations.

This brings us to an important discussion—how India can leverage AI for real-world impact.

Why India Needs AI Growth in Regional Languages

For AI to benefit India, it must be accessible to all linguistic and regional groups. One major hurdle? Most AI models are trained on English, leaving out millions of non-English speakers.

Nilekani advocates for AI4Bharat, an IIT Madras initiative focused on building open-source AI for Indian languages. This project aims to:

✔ Develop AI models for regional Indian languages.
✔ Ensure AI inclusivity for people who don’t speak English.
✔ Empower businesses, education, and governance through AI-driven solutions in local languages.

With India’s diverse linguistic landscape, investing in AI for vernacular languages is crucial for inclusive digital growth.

Conclusion: The Future of AI Lies in Transparency & Practical Applications

The future of AI isn’t just about creating bigger and more powerful models—it’s about making AI practical, accessible, and transparent.

Key Takeaways:

✅ AI secrecy is slowing down innovation.
✅ China’s DeepSeek is proving that open-source AI can change the game.
✅ AI models are commodities, and their real impact lies in how they are used.
✅ India must invest in AI for regional languages to ensure digital inclusion.
✅ The future of AI is in practical applications, not just model size.

By focusing on openness, inclusivity, and real-world use cases, we can harness AI for meaningful impact—just as Nandan Nilekani envisions