What would it look like to apply core open source principles—permissive licensing, transparent training data and weights, and most importantly, community contributions – to AI models themselves?

While several well-known organizations have released open models, many of these aren’t truly open source. They come with restrictions that make them hard to fully understand or build upon. For enterprises, that’s a significant hurdle.

To confidently adapt and fine-tune models for real-world use cases, organizations need complete visibility into how a model was trained, what data it used, and who contributed to it.

OpenShift AI

At Red Hat Summit 2023, Red Hat launched OpenShift AI, a scalable, optimized platform for running AI workloads. While it’s not about delivering models directly, it lays the groundwork for something bigger. Today, Red Hat is going further: it is bringing open-source to the models themselves.

In partnership with IBM Research, they’re open-sourcing several language and code-assist models. Even more exciting: they’re introducing InstructLab, a new open-source project that allows anyone to improve and expand an LLM through a simple interface—just like contributing code via pull requests in traditional open-source projects.

Instead of forking and isolating a model, InstructLab invites global contributions—new knowledge, new capabilities—that can be merged into future versions. You don’t need to be a data scientist to participate.

Domain experts, developers, and data scientists alike can contribute in ways that are meaningful and accessible. That’s powerful—for the community and for enterprise adoption.

RHEL AI

This vision is coming to life through RHEL AI, which combines the strength of Red Hat Enterprise Linux (now with the new image mode), open-source licensed Granite models, and a fully supported distribution of InstructLab. Together, they extend the impact of open source deep into the world of AI, making collaboration and model development feel just like any other open project.

AI innovation shouldn’t be reserved for organizations with vast resources and deep technical teams. Everyone, from developers and IT operations to business teams, should have a path to contribute. That’s the promise of InstructLab and RHEL AI: democratizing AI through the open-source model.

AI for everyone

This is the future of Red Hat’s AI strategy. It’s rooted in the same philosophy that helped us bring open source to Linux, Kubernetes, and hybrid cloud.

Now, they’re doing it for AI. Because AI should be built in the open—and belong to everyone.

Are you eager to know more about this topic? Contact us for a no-obligation appointment.

Do you want to know more about our services? Discover them here.