Why some insiders believe OSI’s new open-source definition falls short

Open Source Initiative’s controversial AI definition sparks debate: The Open Source Initiative (OSI) has proposed a new definition for Open Source AI, triggering significant criticism and concerns within the open source community.

Key points of contention:

• The proposed definition includes access to training datasets as part of the “source” for AI models, which some argue goes beyond traditional open source principles.
• Critics, including Bruce Perens, the original author of the Open Source Definition (OSD), have voiced strong opposition to the OSI’s approach.
• The OSI has been accused of censorship and authoritarianism in managing the discussion around the new definition.

Community reactions and concerns:

• Bruce Perens, OSI founder, has publicly criticized the organization’s efforts to redefine open source AI, highlighting a disconnect between the OSI’s current leadership and its founding principles.
• There are fears that the new definition could effectively turn open source into freeware, potentially undermining decades of established open source practices.
• Critics argue that the OSI lacks the necessary expertise and experience to tackle such a significant redefinition of open source principles.

Procedural criticisms:

• The OSI’s process for developing the new definition has been called into question, with accusations of lack of transparency and community involvement.
• The use of private repositories for discussion, rather than open forums, has raised concerns about the openness of the process itself.

Potential implications:

• The proposed definition could have far-reaching consequences for the open source community, particularly as AI becomes increasingly integrated into software development.
• There are concerns that an ineffective or unimplementable Open Source AI definition could disadvantage truly open projects in favor of commercial, opaque offerings.

Alternative proposals:

• A community-led initiative, hosted on GitHub as OSAID-WIP, has proposed a compromise definition that aims to address some of the most contentious issues in the OSI’s proposal.
• This alternative seeks to eliminate problematic data categories, such as proprietary data available for a fee and non-public data, which have been major sticking points in the OSI’s definition.

Broader context: The debate over the Open Source AI definition reflects larger tensions in the tech industry as AI rapidly evolves and becomes more prevalent in software development. It highlights the challenges of adapting long-standing principles of open source to new technologies and the potential for conflicts between different stakeholders in the open source ecosystem.

Looking ahead: As the OSI moves towards finalizing its Open Source AI definition, with a planned 1.0 release in October 2024, the community faces a critical juncture. The outcome of this debate could significantly shape the future of open source in the AI era, potentially redefining the relationship between open source principles and emerging technologies.