(Credit: Laborant/Shutterstock.com)
IBM and software subsidiary Red Hat have created a new open-source platform for AI development called InstructLab, the company announced Tuesday. The new platform is designed to allow anyone to contribute to large language models (LLMs) and increase access to existing models, according to a statement.
IBM has open-sourced some of its Granite AI models, including its Granite 7B model, for InstructLab users. Red Hat is also launching an Enterprise Linux AI platform (RHEL AI), which also uses open-sourced Granite models and InstructLab tech.
Red Hat SVP and Chief Product Officer Ashesh Badani says in a statement that both InstructLab and RHEL AI will help "lower many of the barriers facing GenAI across the hybrid cloud, from limited data science skills to the sheer resources required."
Truly open-source AI tools aren't too common among the biggest AI firms. Google's Gemini is a closed, proprietary model, as is Microsoft's Copilot, which was built with some open-source tech. Microsoft-backed OpenAI has also kept ChatGPT closed off to the public, despite what its name might suggest. And some so-called "open-source" AI models aren't as open and available as their creators claim, like Meta's Llama 2 model.
IBM's push for open-source AI developer platforms comes as some tech organizations, like Firefox developer Mozilla, take action to support open-source AI. Back in March, Mozilla and a group of nearly 50 nonprofit organizations, academic researchers, and AI firms wrote an open letter to the US Department of Commerce, asking the feds to embrace truly open and transparent AI models so that the landscape doesn't become dominated by a small handful of for-profit firms.