Meta is facing criticism for labeling its AI models as “open source.” Stefano Maffulli, head of the Open Source Initiative, accuses the company of diluting the term and confusing users. According to Maffulli, Meta’s Llama models do not meet the criteria for true open-source software, Richard Waters reports for the Financial Times. Meta defends its position, arguing that existing open source definitions are inadequate for modern AI models. Critics point out that while Meta releases the weights of its models, it withholds crucial technical details. This prevents the free development typical of classic open-source software. The Open Source Initiative plans to publish its own definition for open-source AI next week.