Open-source AI must reveal its training data, per new OSI definition
The Open Source Initiative (OSI) has released its official definition of 'open' artificial intelligence. For an AI system to be considered truly open source, it must provide the complete code used to build and run the AI, access to details about the data used to train the AI so others can understand and recreate it, and the settings and weights from the training. The definition directly challenges Meta's Llama, which has restrictions on commercial use and does not provide access to training data. Meta claims that it restricts access to its training data due to safety concerns, but critics say the company is just minimizing its legal liability and safeguarding its competitive advantage.