Meta under fire for 'polluting' open-source

Meta has faced criticism from the Open Source Initiative (OSI) for its use of the term "open-source" to describe its Llama family of AI models. OSI's head, Stefano Maffulli, argues that this usage is misleading and could damage the integrity of the open-source concept, especially as regulatory bodies like the European Commission seek to promote genuine open-source technologies. While Llama has been downloaded over 400 million times and is seen as a competitor to proprietary models like OpenAI's GPT-4, it does not meet the full criteria of open-source, limiting user experimentation and adaptation. Meta claims its models contribute to AI innovation and acknowledges the complexities of defining open-source in the context of AI. However, critics point out that the Llama models do not allow for the same level of transparency and development as traditional open-source software. Other companies have begun to use terms like "open weight" to describe similar models, emphasizing the need for greater transparency in AI development. The OSI plans to publish a new definition of open-source AI, which will call for more disclosure from developers regarding their models and training data. Maffulli warns that if companies like Meta redefine open-source for their benefit, it could undermine efforts to establish true open-source standards in technology.

- Meta's use of "open-source" for its Llama models has been criticized as misleading.

- The Open Source Initiative emphasizes the need for transparency in AI model development.

- Llama models, while popular, do not meet full open-source criteria.

- Other tech companies are adopting terms like "open weight" to describe limited-access models.

- The OSI plans to publish a new definition of open-source AI to promote better practices.