Meta and Microsoft have launched Llama 2, a state-of-the-art, open source, and versatile AI language model designed for commercial and research use. This model, an upgrade from its predecessor, prioritizes responsible AI use with a more rigorous testing and transparency plan to tackle potential issues. To prevent misuse, an acceptable use policy has also been implemented, along with a responsible use guide -- Engadget reported.
For the first time, Meta is offering pre-trained and conversation-based versions of Llama 2 for free. Microsoft, in partnership, is making the model available via the Azure AI catalog and other cloud tools. It will also run directly on Windows PCs and through third-party providers like Amazon Web Services and Hugging Face.
While the initial AI model was exclusive to academics and researchers, Llama 2 allows businesses to tailor the technology for their needs, including chatbots and image generators. Open source nature of the code will enable third-party scrutiny for biases and inaccuracies. This approach contrasts with rivals like OpenAI's GPT-4, which are often restricted due to subscription or licensing income.
Despite the benefits, there are concerns that large language AI models could potentially be misused by hackers or for spreading misinformation. These fears have led some industry leaders to call for a pause on experimentation to address ethical and safety issues, and for legislation holding AI creators accountable for harmful content.
For Microsoft, the Llama 2 release signifies another effort to maintain its competitive edge in AI against rivals like Google. By collaborating with Meta, Microsoft provides more options for its business customers, especially those looking to fine-tune AI models to their specific requirements.
The new way to do social PR and build relationships with media, influencers and brands.
Explore Glyph →Add your listing
Claim this profile
