What are actually the biggest advantages of open source AI and why is there so much demand for it? These questions were answered in a conversation with Kevin Dubois from Red Hat.
At Kubecon 2025, ITdaily spoke with Kevin Dubois, Developer Advocate at Red Hat. He clarifies the impact of open source AI and how companies, developers, and organizations can benefit from this technology. Additionally, he shares his vision on the advantages of open source AI and the importance of transparency in model training.
What is Open Source AI?
Dubois explains that open source AI is much more than just getting access to free software. “Open source AI means we can download models for free, inspect the training and data, and contribute to the models. It’s not just about using the models, but also about the ability to improve and contribute to them,” he says. This allows companies to use AI without having to worry about hidden costs or restrictions.
The workings of commercial models are often not visible to users. Their training and data are also difficult or impossible to trace. Open source AI offers complete transparency here. “Many commercial models have no insight into their training and data, and that’s a problem for organizations, and may even be illegal,” Kevin explains.
Benefits for Developers
One of the biggest advantages of open source AI for developers is the ability to run models locally. This gives developers full control over their data without the need to use commercial AI solutions that often come with restrictive license terms.
“With open source models, you can run them locally, without worrying about licenses or internet access,” Kevin says. “That’s especially useful when working with sensitive data, because you have full control over where your data is stored.”
Red Hat offers developers tools like Podman AI Lab, which allows them to easily run AI models locally. “With these tools, developers can run AI models without relying on external servers, which not only reduces costs but also improves feedback speed,” says Dubois. Additionally, the company launched Konveyor AI and a renewed Developer Hub.
Transparency and Data in AI
Another important theme in open source AI is transparency, especially regarding model training and the data used. Many commercial AI models are closed and provide no insight into the data they were trained on. This can lead to problems in terms of privacy and copyright.
“If I ask a model where the data it was trained on came from, the model should be able to answer that,” says Dubois. Red Hat is therefore working with parent company IBM on the Granite models, which are more transparent in the data used and the training process.
Misconceptions About Open Source AI
He emphasizes that there are many misconceptions about open source AI. A common misconception is “that open source AI is always free.”
“People often think that open source always means ‘free’, but it goes beyond that. Open source means you have access to the code and the ability to contribute, but it doesn’t always mean it’s without cost,” he explains. The licenses that come with open source software can have restrictions. It’s important for developers to be aware of this.
The Future of Open Source AI
According to Dubois, open source AI has the potential to accelerate innovation in the AI space. “Open source means open collaboration, and that accelerates innovation,” he says. Red Hat believes that the future of AI will lie in open source, just as open source operating systems and containers like Kubernetes have made their mark on the tech world in recent years.
Kevin Dubois, Developer Advocate at Red Hat“We need to open up the world of AI so that everyone can contribute to and benefit from this technology”
“We are already seeing that the adoption of open source AI is happening faster than with commercial models, as we have seen with other open source technologies,” says Dubois. Promoting open standards and collaboration ensures a faster evolution of the technology.
Open Source vs. Commercial AI
When it comes to the competition between open source AI and commercial companies like OpenAI, Dubois says that open source is currently not really able to make the same investments as those of large commercial companies. “In an initial phase, it’s more difficult to compete with companies that invest heavily in commercial AI solutions,” he says. “But ultimately, open source will prove itself, as we have seen with other technologies.”
The emergence of open standards, such as Model Context Protocol (MCP) from Anthropic, is an example of how open source AI can evolve faster than commercial models. “MCP is an open source initiative that accelerates the adoption of generative AI and enables developers to add new capabilities to large language models,” says Dubois.
In Full Growth
Open source AI will play an important role in the future of artificial intelligence. It offers developers the freedom to run models locally, ensures transparency in model training, and promotes open collaboration.
While commercial AI models currently still dominate the market, he believes that open source AI will prove itself in the long term through its speed of innovation and the open standards it supports. “We need to open up the world of AI so that everyone can contribute to and benefit from this technology,” Dubois concludes.