Week 8: Open Source AI Reflection

Takeaways from Nick Vidal from OSI

This week, we had Nick Vidal from the Open Source Initiative (OSI) present about the Open Source AI Definition. The presentation deepened my understanding of what truly constitutes “open source” in the AI space.

I was particularly struck by the comprehensive requirements outlined in the Open Source AI Definition. It’s not just about making code available - a truly open source AI system requires transparency across multiple dimensions:

1. Data Information: Detailed information about training data, its provenance, and selection methodologies
2. Code: Complete source code used to train and run the system
3. Parameters: Model parameters like weights and configuration settings

The document makes clear that all these elements must be available under OSI-approved terms to be considered properly “open source.”

I also learned about “openwashing” - the practice of misleadingly labeling products or projects as open source when they don’t meet the fundamental freedoms required by the OSI definition. This seems particularly prevalent in the AI space where companies might release limited parts of their systems while keeping critical components proprietary.

Future of Open Source AI

Looking forward, I see several important trends for open source AI:

1. Community-driven innovation: Just as with traditional open source, community collaboration will likely drive rapid innovation in open AI models
2. Legal frameworks evolution: New licensing approaches specific to AI will need to develop, as noted by Heather Meeker: “permissive open source licenses specifically for ML models don’t really exist yet”
3. Balancing openness and responsibility: The community will need to navigate how to maintain openness while preventing harmful applications
4. Democratized access: Open source AI has potential to reduce barriers to AI development, allowing more diverse participants

Open Source in Finance

While we didn’t have the scheduled speaker from FINOS (Fintech Open Source Foundation), I’ve been reflecting on open source’s role in finance. Financial institutions have historically been cautious about open source due to regulatory and security concerns. However, there’s been growing recognition that collaborative development can address common industry challenges.

Open source in finance offers standardization benefits while allowing institutions to customize where needed. Projects like FINOS demonstrate how traditionally closed industries can benefit from collaborative development while maintaining appropriate security and compliance controls.

FOSS Figure Research: Larry Wall

Since our speaker couldn’t make it, we explored famous FOSS figures. I researched Larry Wall, an influential American computer programmer, linguist, and author.

His major accomplishments include:

• Creating the Perl programming language
• Developing the Unix “patch” program
• Founding the Perl Foundation to support programming communities
• Authoring the “rn” Usenet client

Interestingly, Wall was initially trained as a linguist and wanted to document unwritten languages in Africa. His programming language Perl reflects his linguistic background, with the name being a biblical reference to the “pearl of great price” (Matthew 13:46). Wall received the first Award for the Advancement of Free Software in 1998.

Team 3 Oppia Contribution Update

Our team has made good progress on our Oppia contributions. Everyone has completed the Contributor License Agreement and the Oppia contributor survey. Each team member has identified and posted at least two issues they’re interested in working on, and we’re moving toward finalizing the specific pull requests we’ll tackle.

We’ve been meeting regularly (three times this week on 3/11, 3/13, and 3/16) to coordinate our efforts and support each other through the contribution process. So far, the collaboration is going smoothly, and we’re all learning a lot about the project’s codebase and contribution workflow.