Does Open Source Make More Sense for Learning in the Age of AI?

The debate between the merits of open source vs proprietary systems has raged for about two decades. But the sudden emergence of generative AI has reignited the debate and dramatically increased the stakes.

Educators the world over have long considered LMS solutions such as Moodle against commercial providers such as Brightspace and Blackboard. The paid-for products often look slightly more modern, while the open source community espouses the benefits of customisability and the lack of vendor lock-in. Yet, the core functionality of an LMS hasn’t changed significantly in years; it’s a mature product category that’s approaching commodity status. The new battleground will be how effectively an LMS can integrate AI, and that’s where proprietary systems may suddenly be at a considerable disadvantage.

The integration of AI will be a game-changer in education, promising to redefine how learning is delivered, experienced, and managed. Central to this transformation for institutions will be the choice of AI technologies to deploy. There are two layers to this choice – the LMS platform, and the AI Large Language Models (LLMs) to be integrated. Just as in the realm of Learning Management Systems, similarly in the world of AI, there are both proprietary and open source solutions available. Choosing a proprietary LMS risks being restricted in your choice of AI technologies. And choosing a commercial LLM vendor risks incurring cost, privacy and suitability challenges. 

Keeping Up With AI

Although most of the mainstream AI headlines have been dominated by the likes of OpenAI (creators of ChatGPT) and Google (creators of Gemini), one of the most notable developments has been the speed and scale of open source AI model availability. Free-to-use models such as LLaMA or Mistral regularly challenge commercial model performance in benchmarks.

The speed of change in AI is unlike anything we’ve seen before. And that’s why vendor tie-in is such a concern in this environment. With new, dramatically more capable models being released at a rapid cadence, using an open source  LMS gives you the ultimate in flexibility – choose when to integrate with a commercial AI or an open source AI – or both. Choosing an open source LMS allows you to respond quickly to changes in AI technology, experiment with technologies that are relevant to your specific needs and give you far greater control over ethical and privacy concerns. 

Cost is also a significant consideration as we look to harness AI to create better learning experiences. Organisations that choose to use commercial AI services will need to carefully monitor the running costs as usage increases. The ability to offload tasks to an open source LLM offers compelling cost control. 

Perhaps the largest area of concern around sticking with a proprietary route is around the ethical and privacy issues, coupled with a lack of transparency. Using commercial AI models raises risks around data leakage, training data bias and a lack of clarity on the training data used and guard rails implemented. In an era where data privacy concerns are paramount, the transparency afforded by open-source AI is a significant boon. Educators and institutions can scrutinise the open source AI features integrated into their open source LMS, understanding how data is processed and ensuring it aligns with ethical standards and privacy regulations. This transparency fosters trust among users, an essential component in the effective adoption of AI technologies in education.

As we move past the early phases figuring out what LLMs can do in the education space, these operational questions will become central to the deployment phase of the technology. The progress of open source AI models alongside their commercial cousins is a very positive development for democratising access to this most powerful technology. Regulators are already cognisant of the need to deploy these technologies in a responsible and transparent way. 

Proprietary LMS and commercial AI that limits choice risks negating the accessibility and inclusivity potential of LLMs. Progressive institutions can’t risk being constrained by commercial providers who might choose to prioritise their agenda over opportunities to develop and deliver enhanced learning outcomes. The academic opportunities offered by fine-tuned LLMs built on open source models that can be easily integrated into open source Learning Management Systems may be the biggest single argument against closed platforms in two decades.

Image: While open source LMS allow for the use of both proprietary and open source LLMs, proprietary LMS may not provide as many options as the technology evolves.

The Future is Open

The emergence of AI will bring about dramatic changes in learning institutions. Those that review their current technology choices through a future-focused lens that prioritises flexibility for an uncertain world would be wise to choose an open source LMS that can embrace both commercial AI and open source AI. A vendor-agnostic approach to AI development significantly amplifies the possibilities in this arena, offering advantages far beyond those found in proprietary software.

Educational institutions that choose to embrace the potential offered by open-source will ultimately create the most effective learning experiences for their students, the most impactful tools for their lecturers and the most ethical platforms for their organisation.