The Open Source Initiative (OSI) has released version 1.0 of its Open Source AI Definition (OSAID), establishing the first industry standard for determining whether an AI system can be considered truly open source. Developed through years of collaboration with academia and industry, the OSAID requires open source AI to provide sufficient information to substantially recreate the model, including details about training data provenance, processing, and licensing. It also grants usage rights such as the freedom to use, modify, and build upon the model.
While the OSI lacks enforcement mechanisms, it plans to flag AI models incorrectly described as “open source” that fail to meet the OSAID criteria. Several major tech companies have used the term for AI releases without fully adhering to open source principles.
One such company is Meta, well known in the AI scene for its Llama models. They are not open source according to the OSAID, because Meta doesn’t provide access to its training data. It also puts restrictions on commercial use. Meta spokesperson Faith Eischen told The Verge:
“There is no single open source AI definition, and defining it is a challenge because previous open source definitions do not encompass the complexities of today’s rapidly advancing AI models.”
The definition’s first version does not comprehensively address issues related to proprietary data licensing or copyright in the context of AI models. To manage the OSAID’s evolution, the OSI has formed a committee tasked with monitoring its application and proposing amendments for future versions as the complex landscape of open source AI continues to evolve.
Sources: TechCrunch, Open Source Initiative