Talk
Intermediate

Open Source in AI: A tour and Analysis

Rejected

Session Description

Open source has made (and helped make) rapid strides in AI in recent times—far more than public perception gives it credit for.
This session takes a tour through key projects, research, and collective efforts that have quietly shaped the field from the ground up. It follows with an analysis of important but unobvious patterns, and a brief tour of what's hindering it from making a full bloom.

We explore:

  • plenty of open source endeavors in a whirlwind tour

  • why open ecosystems matter deeply to the future of AI

  • how current open source AI is shaped—often steered by corporate priorities

  • the tendency for contributions to cluster around hype-driven moments

    • the flood of LLM-centric and LLM-wrapper projects is a prime example

  • how important but untrendy work often goes unnoticed

  • ....and more miscellaneous patterns

  • important gaps in the open-source AI ecosystem

  • the wealth of open research, datasets, and prototypes that remain underutilized

And we reflect on why open source, by itself, falls short of its full potential:

  • limited access to essential hardware

  • uneven distribution of expertise and context

  • lack of supportive policy and governance frameworks

    • IP and copyright

    • lack of focus on non-mainstream research

This is more than a survey — it’s a critical mapping of what’s working, what’s missing, and what we could co-create next.

Key Takeaways

  • Open source has quietly driven major advances in AI—far more than public perception reflects—through a wide range of community projects, tools, and research.

  • Open ecosystems are essential to AI’s future: enabling transparency, equitable prosperity & access, and shared progress.

  • Patterns in open source AI reveal:

    • corporate influence shaping much of the development

    • hype cycles steering attention and effort

    • crucial but untrendy work often overlooked

    • vast open research, data, and tools left underused

  • Open source alone isn’t enough. Bottlenecks persist in:

    • hardware access

    • knowledge and skill distribution

    • policy and regulation

References

Session Categories

Community
survey of a FOSS ecosystem
open source in AI
Knowledge Commons (Open Hardware, Open Science, Open Data etc.)
Other
open source in AI
Which track are you applying for?
Main track

Speakers

Priyanshu Mishra
Freelance ML Engineer and Consultant
https://priy.me
Priyanshu Mishra

Reviews

0 %
Approvability
0
Approvals
1
Rejections
1
Not Sure

A thorough proposal, but doing justice to the matter requires significant experience/expertise/research. The proposer has listed several references, but none of them point to their personal experience/hands-on knowledge on the topic.

Reviewer #1
Not Sure

For future submissions, we recommend that you either present on a project to which you are a direct contributor or provide more concrete examples and references to demonstrate your personal experience and expertise in the topic. This will help the program committee better assess the value of your talk to the IndiaFOSS community.

Reviewer #2
Rejected