Beyond Open Models: Building an Open AI Commons for India

AI is rapidly becoming one of the most consequential technologies in the world, yet much of the AI stack remains concentrated in a handful of companies, platforms and closed ecosystems. Open-source models are an important step forward—but models alone do not create an open AI ecosystem.

This talk explores what it would take to build an Open AI Commons for India: a community-driven ecosystem spanning open models, datasets, tools, research, compute, documentation, education and governance.

Drawing from the journey of Bharat AI Mission, an open-source AI community initiative, the session will examine the opportunities and challenges of building AI as a shared commons rather than simply consuming AI as a service.

We will explore questions such as: What should actually be open? How can developers and researchers contribute beyond writing model code? How can universities, startups, industry and communities collaborate? And what can the FOSS movement teach the emerging AI ecosystem about openness, collaboration and community governance?

The session will conclude with practical ways for developers, students, researchers and FOSS communities to start contributing to India's open AI ecosystem.

1. Why open-source AI requires much more than simply releasing model weights or code.

2. The different layers of an Open AI Commons: models, data, tools, research, compute, documentation and community.

3. How FOSS principles can help shape a more open, inclusive and community-driven AI ecosystem.

4. How students, developers, researchers and organisations can contribute to open AI beyond building applications.

5. Practical pathways to participate in and build India's emerging open-source AI ecosystem.