As Git considers a voting procedure to determine AI (LLM) usage policies, the open-source community is exploring new democratic governance methods to address technological changes.
Imagine this. In a core tool used by tens of millions of developers worldwide, rules need to be set on how artificial intelligence (AI) should be used for writing code. Instead of a small number of managers making this crucial decision behind closed doors, how can it be decided through a ‘vote’ by the entire community? Recently, an interesting movement has been drawing attention in the global developer community. Git, the standard version control system, has started considering a voting approach to establish policies related to large language models (LLMs, AI technology that learns massive text to write like a human and write code) [Source 1, Source 8].
Why Is This Important?
In the software development world, ‘Git’ is like an official registry that manages a building’s blueprints. Anyone uses this registry to fix code, add new features, and collaborate with people around the world. It is essential foundational knowledge that anyone learning computer science or coding must know [Source 2]. To put it metaphorically, it is like a ‘national library’ that records and manages revision histories without a single error when numerous chefs around the world collaborate on a single giant recipe book.
However, with the explosive growth of generative AI in recent years, the global development field has fallen into significant confusion. This is because there were no clear standards on whether AI-generated code could be used as-is, whether copyright issues existed in AI-created code, or to what extent AI usage should be permitted in the contribution process. In fact, some open-source projects (such as yt-dlp, famous as a video download tool) require contributors to check a checklist confirming “I have read the AI usage policy (NO AI / NO LLM POLICY) and understand that violating it may result in being banned from the repository” when submitting code or issues [Source 3].
In this situation, the fact that a core infrastructure project like Git has brought out a democratic process like ‘voting’ to set AI policy means that it can serve as a standard precedent for numerous open-source projects to establish rules in the AI era [Source 1, Source 8].
Easy to Understand: A Democratic Experiment in Rule-Making
Let’s put it this way. Imagine residents gathering in an apartment complex to create guidelines for raising pets. If the management office director unilaterally sets the rules, residents will push back. Instead, if rules are set through a vote involving all households, everyone is likely to accept the result.
What Git is trying to do is similar. When the pace of technological development is extremely fast and stances are divided, it is a process of contemplating how community members should reach a consensus on the political and ethical impacts or bias issues that AI will bring [Source 5]. The open-source ecosystem has originally grown through voluntary participation and discussion, but in front of the giant wave of AI, it has seriously put institutional devices (voting processes) to transparently poll and decide community intentions on the table [Source 1, Source 8].
From a broader perspective, this closely resembles the process by which a democratic society finds appropriate policies and values in the face of new technologies. Just as society members must analyze potential outcomes based on various indicators and seek better directions, the development society is also experimenting with systematic decision-making structures amidst technological changes [Source 15].
Current Situation: How Far Has It Come?
The software development ecosystem is currently passing through a period of upheaval. Amidst growing voices considering policy changes of large code hosting platforms or alternatives to centralized services [Source 10], there are pros and cons regarding AI receptivity within the Git community, which forms the basis of development tools.
On one hand, some argue that productivity should be maximized by utilizing AI, while others demand strict control or limitations due to code quality, license contamination, and security risks. The discussion to establish official policies through clear voting procedures rather than forcing conflicts with power or leaving them ambiguous has now moved into full swing [Source 1, Source 8].
What Will Happen Next?
What voting method Git will ultimately adopt and what LLM policies it will finalize are drawing intense attention from the global open-source community. If policy decision-making through voting successfully takes root, many other open-source projects are likely to adopt similar governance models.
Developers will be able to clearly know under transparently disclosed rules how projects they contribute to will review and accept AI-written code in the future. Git is showing the lesson that as technology develops, the way rules are made must also evolve together [Source 1, Source 8].
References
- Git considers voting to determine its LLM policy
- We Need To Stop Lying About Git - YouTube
- ok.ru not downloading any video · Issue #17585 · yt-dlp/yt-dlp · GitHub
- Hidden Persuaders: LLMs’ Political Leaning and Their Influence on…
- Welcome to LWN.net [LWN.net]
- GitHub is sinking – David Bushell – Web Dev (UK)
- LLMs’ Potential Influences on Our Democracy: Challenges and…
- Sole decision by top management
- Voting process
- Automated AI algorithm judgment
- NO AI / NO LLM POLICY
- AI automatic generation of all code
- Complete ban on commercial use
- To speed up software building
- To find appropriate direction amidst social and technological changes
- To reduce server costs