Meta Platforms has confirmed that it will continue to release artificial intelligence tools for approved researchers, despite claims that its latest large language model has been leaked to unauthorized users. In a statement on Monday, Meta noted that "although the model is not available to everyone and some have attempted to circumvent the approval process, we believe that our current release strategy allows us to balance responsibility and openness."
Meta, the parent company of Facebook, has a major research unit in artificial intelligence and released a language model called "LLaMA" last month. The company stated that the model is capable of generating human-like conversation, a feature developed by OpenAI, the creator of the ChatGPT chatbot, and Alphabet, the parent company of Google, while utilizing much less computing power.
Meta offers its tools to researchers and other entities associated with government, civil society, and academia under a non-commercial license after a vetting process. Last week, users in an online forum claimed they made the model available for download.
