With the focus on open education and open technologies during Open Education Week, which runs from 3 to 7 March 2025, let’s explore some ways in which open content is being harnessed by Artificial Intelligence (AI) models to deliver quality education to some of Africa’s youth.
Is AI the tool Open Educational Resources (OER) practitioners have been waiting for? One of the biggest reasons educators have been slow to embrace OER is not because of the value proposition; they have always appreciated the promise of free access to quality resources. Rather, it is the time required to search for, evaluate, adapt, and attribute OER that has been a key deterrent to full adoption. In our busy schedules, adapting OER was often seen as just too time consuming. Now, AI can streamline and accelerate processing of OER.
Moreover, when mediated by educators, AI can do much more than just adapt resources. It can be used at most stages of designing and delivering education. A recent academic paper provides evidence that the education community has been particularly receptive to using AI tools – in this instance Claude AI – to mediate education instruction.
A notable example from Africa is YYeni AI, based in Windhoek, Namibia. This startup, led by two young entrepreneurial Namibians, employ various AI tools to enhance secondary school education.
At A. Shipena Secondary School in Katutura, the YYeni AI is used to administer diagnostic tests that assess the current understanding of grade 10 Mathematics and Physics students.
The YYeni AI generates questions and, based on the students' responses, develops individualized training curricula that are tailored to their strengths and address their weaknesses. It then creates unique learning pathways for each student, using learning materials developed from openly licensed content.
The content is accessed via laptops, tablets, or smart phones, and students are regularly tested as they progress through the materials. The subject content includes text, images, and some videos.
Due to limited space, the students in the AI pilot class have no dedicated classroom, nor do they have a teacher. Instead, they each have a simple desk arranged in rows on the school's assembly hall stage, with curtains drawn to allow other activities to continue in the hall. A YYeni AI facilitator moves around this makeshift space, answering both technical and educational queries as they arise. An Internet router has been set up for their dedicated use. Learners progress through their individualized programmes at their own pace. Breaks are organized to coincide with regular classes. In a resource-scarce environment, openly licensed content and AI tools are being harnessed together to offer something special.
This initiative in Namibia is just one example of several AI-driven education experiments taking place across Africa. Interest in AI for open education is growing, particularly because of its potential to create adaptive learning environments. These environments use AI algorithms to continuously assess and adjust the learning path based on the learner's progress and preferences. This ensures that each student receives a tailored educational experience that maximizes their potential.
Moreover, AI can enhance accessibility in open education by providing tools that support diverse learning needs. For example, AI-driven language translation and speech recognition technologies can help non-native speakers and learners with disabilities to access and engage with educational content more effectively. Publicly accessible AI driven translators include Google Translate and Microsoft Translator. Google’s AI Tutor and Microsoft’s Immersive Reader provide interactive experiences where learners can practice real world conversations with real-time feedback.
However, both open and proprietary communities have raised concerns about the sources AI relies on to generate and adapt content. Who owns the data these models collect, and are there inherent biases in these datasets that are being perpetuated? From an African perspective, does data from the Global North adequately reflect African contexts? Additionally, from an educational standpoint, is the way AI creates learning pathways pedagogically sound and mindful of the educator-student dynamic?
Of particular concern to OER practitioners in Africa is whether the sources being used are properly attributed. While they encourage the sharing of their resources, they also seek recognition for their contributions. Evidence suggests that many AI tools often overlook this crucial attribution step when remixing resources into new learning materials.
Also, AI can create content using a variety of sources, including open content, but it does not exclusively use openly licensed content unless specifically instructed to do so. The sources of content depend on what data the AI has been trained on, which can include a mix of OER, publicly available information, and proprietary content.
To determine where the content originated, some AI tools and platforms provide mechanisms to trace the provenance of the information used. For instance, certain AI content creation tools can track and report the sources of the data they use. One example is Microsoft Copilot that provides references with navigable links back to the original sources at the end of generating an output. This helps ensure transparency and allows users to verify the authenticity and origin of the content.
While AI tools seem to offer some long-awaited solutions for OER practitioners, they are also transforming the creation and future use of openly licensed content. If we are not vigilant, the core values of the open community – equitable access to quality education for all – could be compromised. It's crucial to advocate for AI tool developers to uphold these values. This includes ensuring ethical data use, promoting equitable access, and galvanizing international collaboration to build a more inclusive educational ecosystem.
Open Education Week takes place from 3-7 March 2025. Find out how you can contribute here: https://oeweek.oeglobal.org
Images courtesy of Andrew Moore (CC BY-SA)
Related articles