Speakers
Description
A year ago, I attended Opportunity Open Source 3.0 as a participant, curious about how open-source projects actually function behind the scenes. Since then, I have gone on to contribute actively across multiple open-source ecosystems, including through Code for GovTech (C4GT), where I worked closely with maintainers and real-world codebases used at scale.
Somewhere in that journey, I noticed a growing pattern: AI coding assistants are being used not to assist contributions, but to replace the thinking behind them. Pull requests are increasingly AI-generated end-to-end — plausible-looking, syntactically correct, and completely disconnected from the project's actual architecture, conventions, or intent. The result is a quiet erosion of PR quality across open-source repositories, one that maintainers are now spending disproportionate time catching and rejecting.
This talk is about that erosion — why it's happening, what it looks like from a contributor's and maintainer's perspective, and how AI-assisted contributions can be brought back to a place where they add real value instead of review overhead. Using real examples from my own contributions and reviews, I will break down the difference between using AI as a research/learning aid versus using it as a submission generator, and share practical checks contributors can run on their own PRs before submitting.
The session is intended for students, first-time contributors, and maintainers who want a grounded, honest look at how AI is reshaping the open-source contribution pipeline — for better and worse.
Attendees will leave with a clearer framework for using AI responsibly in their contribution workflow, and maintainers will get language and criteria for identifying and addressing low-effort AI-generated submissions.
Any other info we should know?
The audience will learn:
• Why AI-generated PRs are increasingly rejected by maintainers.
• The difference between using AI to learn vs. using AI to submit.
• Real patterns of low-quality AI contributions seen across projects.
• A practical self-check framework before submitting an AI-assisted PR.
• Lessons from contributing through C4GT and other open-source ecosystems.
The talk is beginner-to-intermediate friendly, grounded in real PR examples rather than theory.
Session author's bio
Pankaj Kumar is an open-source contributor who began his journey as a participant at Opportunity Open Source 3.0. Since then, he has contributed to multiple open-source projects, including through Code for GovTech (C4GT), and has developed a strong interest in the intersection of AI tooling and contribution quality. He is passionate about helping the community use AI responsibly rather than as a shortcut that quietly undermines the value of open-source contributions.
| Agree to Privacy Policy and Notice | I agree |
|---|---|
| In Person Attendance | In-person |
| Level of Difficulty | Beginner |
| Please confirm that there are included headshots of all speakers in their profiles | Yes |