Facebook’s parent company, Meta, has released Llama 2, its large language model for generative AI, for commercial use. The code will be available freely under a customized partial open-source license. This move aims to attract developers and provide an alternative to OpenAI and other similar platforms. However, Meta will have to address the potential issues arising from the content generated by these programs. While Meta won’t charge for using Llama 2 directly, it plans to sell the program as a paid hosted service through cloud computing companies. Microsoft will be the first to offer this service, followed by Amazon and others. Meta believes in an open approach to AI model development, especially in the generative space, as it allows for widespread benefits. Despite being a big company itself, Meta expects to benefit from the advancements made by others using Llama 2. The company’s VP of generative AI, Ahmad Al-Dahle, highlights the versatility of large language models in generating text. Meta has an acceptable use policy in place to regulate commercial usage of Llama 2, prohibiting illegal activities and harassment. The model underwent rigorous testing, including adversarial prompts, to refine its performance. The first version of Llama was released in February for research purposes and received over 100,000 requests.
