15 October 2025, Bamako, Mali
Umaru of Mali had a sleepless night. He’d spent most of the evening trying to learn about the intricacies of international taxation so as to best prepare himself for an internal job promotion. He used the YeeHah learning platform. He was struggling especially on the double tax treaties that exist between the African Union and the European Union.
The next morning, as Umaru was boarding the solar-powered city rail in Bamako, he put on his headphones and accessed YeeHah. The algorithm powering YeeHah immediately started to provide case studies of how tax impacts trade between Africa and Europe, with expert tax practitioners sharing case examples. Umaru closed his eyes and immersed himself in the world of tax with a quiet smile of satisfaction on his face as he began to understand the intricacies of international trade.
YeeHah is an open and dynamic learning algorithm-driven platform that’s free – the ‘free’ stands not for the absence of payment, but for freedom. The freedom to create, curate and distribute content through an open-source algorithm which is highly adaptive to individual learners and provides them with knowledge and learning in a manner that is appropriate to their needs at any point in time.
Dealing with a generational challenge
In a world of rising economic inequality, technology has the potential to exacerbate the widening disparity. The pace of technological advancements throws into question the future viability of jobs and employability in their current form. Whilst there is the potential and opportunity for new jobs and roles, the fears of significant unemployment and social dislocation due to the potential inability of the human race to match technological scale and advancement remain unabated.
Greater social mobility remains of the strongest tools at our disposal to combat inequality. We need to consider how we can best improve the quality of opportunities and life chances for all. These opportunities start from educational attainment progressing on to professional development and ultimately life fulfilment. We will need to consider how we can best narrow life outcomes between those starting with disadvantaged backgrounds and their affluent peers.
Education has the potential to have a transformative impact on social mobility and be a leveller of inequality. Access to education comes through institutions such as schools and increasingly also through digital platforms which, by their design, present few artificial barriers to attainment.
There’s a risk, however, that schools and educators may not have the necessary tools to provide the type of learning and education support desired by learners. In a world beset with technological inequality, further problems arise when those with access to technology have better educational attainment than those without.
Here’s a proposal on how we can resolve this particular challenge.
The power of open algorithms
Artificial intelligence (AI) has the revolutionary ability to hand educators the tools they need to support, guide and inspire learners the world over with ease and efficiency. How algorithms or AI differ from traditional adaptive learning systems is their ability to continuously develop deeper insights and be able to adjust their educational content delivery based on reasoning and updated decision-making. However, there remains a risk that with a proprietary approach towards building algorithms that power AI, the status quo of social and technological inequality remains broadly entrenched.
This is where a collaborative, multi-stakeholder, multinational effort to build common and open standards in algorithms pertaining to education could make a positive difference to the world we live in.
In this world, effort should be applied towards the building of an open-sourced algorithm that has the ability to remain adaptive to learner needs. Adaptive learning technologies already exist presently. However comprehensive adaptive learning algorithms can apply the powers of deep learning to best understand the native learning habits and styles of learners. They will also be able to deliver the content to learners in ways they desire, at the time they require, and in an appropriate and contextualised manner.
This intuitive hyper-adaptive (I.H.A) algorithm will create a better understanding of the profile of an individual learner and better predict their behavioural patterns and cognitive abilities from insights continuously drawn from vast numbers of users adopting the algorithm. The scale of usage will enhance the predictive and intuitive abilities of the algorithm.
A key challenge with AI is the possibility of bias creeping into the algorithm that may result in the erosion of trust. This is why any open algorithm developed through a global network of partners will ensure a greater diversity of data and open-source development will help negate the possibility of bias.
The transformative impacts of AI
Individuals have different approaches and attitudes to learning. Our individual cognitive abilities and styles shape our approach to learning and the way we consume content. Even the way we learn different subjects can be different. For instance, we might learn about history through reading. We might learn mathematics by carrying out maths exercises (or by ‘doing’), or through reflection. We might learn physical geography through observation, video and audio. These learning preferences or styles are not static and have contextual differences that need to be reflected in the way courses are designed and created.
However the myriad of factors that impact learning preferences (including culturally-specific and contextual backgrounds) mean that it can be very difficult for an educator to cater to the individual and unique needs of the many. Attributional diversity can often be greater than the similarities amongst learners. This could mean opportunities are missed to create a community of learners who are significantly more motivated, engaged and inspired.
This is where the power of algorithms, machine learning and AI can step in to provide an indispensable pillar of support to educators to improve educational attainment and the performance of their students.
Through the development of a heuristic model underpinned by data collection on how individual learners prefer to consume educational content, it will be possible for algorithms to build individualised learning profiles. These intuitive hype- adaptive (I.H.A) learning systems will continue to learn about how individuals educate themselves; understand the commonalities across different individuals who share similar behavioural characteristics; and continuously refine and adapt the way educational content is made available to them.
The machine learning abilities offered by AI allow for not just merely a descriptive set of analytics which provide a current state view of the world but more crucially predictive as well as prescriptive analytics. Predictive and prescriptive analytics will provide an analysis on how other learners in aggregate are learning different subject areas and predict and prescribe and suggest content and educational material offered to learners in a way that will be both impactful and relevant.
The data insights and models developed by the AI which forms the basis of an intuitive hyper-adaptive learning algorithm and platform will ultimately help drive both effective educational attainment as well as a continuous improvement of the learning interface.
This tailoring of education is based on the algorithms’ abilities to not just pre-empt or detect an individual’s learning styles but also to anticipate their on-going educational needs as they go through their educational journey. With advances made in the volume of relevant data collected as well as enhanced computing power, algorithms will be able to rapidly create a detailed model of student profiles based on deeper understanding and interpretation of learner’s cognitive levels, their learning approaches and affective states.
Educators can reap significant benefits by providing knowledge in ways that are best positioned for their students. There is significant research evidence[i] to support the notion that relevant instructional and educational interventions that are matched to learning preferences will ultimately lead to enhanced outcomes.
How educational content could look in the future
In order to support hyper-personalisation to learners through a dynamic learning interface, it will also be crucial to consider how educational content is developed and delivered.
In order for an algorithm to serve the needs of learners most effectively, it will be important to have content in an omni-format approach across multiple cognitive levels. Educators will need to consider developing content that can be delivered in multiple formats (interactive, audio/video-based, text-based – i.e. omni-format) so that the algorithm can serve the right content at the right time in the format learners need it in.
The learning behaviours on a dynamic learning platform can only be supported with the availability of the relevant content set in different forms and built at different levels. Over time, the algorithm will be able to identify an individual learner’s optimal path to educational attainment and achievement and provide educators with the time and guidance required to enhance their own teaching.
Through this approach, for educators have the potential to make a positive impact on student learning outcomes at scale. It will also ensure that all students are able to achieve mastery of similar learning outcomes but through an adaptive and intuitive process that takes into accounts students learning preferences and styles.
The creation of the same content in an omni-format manner will mean that learners will not see all versions of the same content as it will be bespoke to their specific needs and demands. These needs are identified through the diagnostic and predictive abilities of the algorithms to best meet learners’ needs.
The algorithm will also be able to model the affective states, cognitive and educational levels of the students as well as their behavioural traits to accurately establish individual profiles. This will help educators shape their own pedagogical approach to their students and also help all learners starting at different levels arrive at the same learning outcome through a hyper-personalised journey of support and engagement. It will also over time allow for educators to build a view of how long it takes for educational attainment to be achieved for the various cohorts of learners and in the process develop more meaningful planning.
These developments will also allow for employers, working closely with educators, to play a role in helping define key educational outcomes and helping deliver the employability promise to learners in a dynamic workplace. Employers can provide details of key employability needs and skills as emerging requirements appear and the relevant content could be developed jointly and subsequently distributed to learners in the ways that will best suit their needs and support them in the workplace.
Delivering broader impacts for all
There is an opportunity for the above education model to drive benefits to wider society globally rather than just specific segments. Whilst other broader factors such as targeted government policy and societal efforts play a big role in supporting enhanced social mobility, access to education and enhancements to educational delivery is a significant tool and enabler of social mobility.
An algorithm delivered through a dynamic learning interface can ultimately help create much more efficient and flexible pathways to support education attainment and can provide educational access to underserved learners.
The development of a relevant algorithm alongside a response learning platform is a significant undertaking and the risk of them remaining proprietary is that it could only further exacerbate the inequality gap and place further constraints on social mobility.
A proprietary algorithm will limit high-quality education to only those with the means to access it. This is why an open-sourced approach will allow various institutions and educators (regardless of their socio-economic backgrounds) to understand the underlying technology, tweak it to meet their requirements and populate it with content that’s relevant to their learners’ needs. This will allow for societal benefits to be delivered at scale. It also allows for the algorithms to be customised to meet cultural and contextual needs and ensure localised learning needs and demands are met.
An open-algorithm, developed through global partnerships and a shared vision, which provides educators with the freedom to create, curate and distribute content to meet their learners’ requirements, has the potential to enhance the overall quality of education delivered, support access to education and attainment, be more efficient, and ultimately improve student outcomes.
Educators can also focus their energy and time towards shaping learners through their craft of teaching and inspiring them towards educational success in a way that is meaningful and personalised. A global delivery of the algorithm with its own dynamic platform will help establish a global community of educators delivering significant local impacts.
Emergent technology has a vital role in ensuring education remains accessible, supports the global attainment of knowledge, paving the path towards greater social mobility, and a more equal world. Ultimately this will allow for progress to be driven in the broadest possible way and allow for high-quality education, underpinned by AI, to deliver greater opportunities for all.
15 October 2035, Addis Ababa, Ethiopia
Umaru, now serving as a senior policy maker at the African Union, is drafting a series of recommendations to help improve governance and business environment across Africa. He is also started sharing his views and knowledge through the YeeHah platform which helps tax administrators, not just in Africa but globally as well, to understand key issues and developments in the area of tax policy and administration. His deep insights have been invaluable for everyone seeking to learn more about tax regimes in Africa and are open to all who seek knowledge.
One thing still perplexed him. Why call this fantastic educational platform YeeHah and what’s behind its name? As he started searching for answers, a smile spread across his face as he started learning about the intuitive hyper-adaptive (I.H.A) algorithmic-driven platform and read out the acronym aloud.
[i] Pane, J. F., Steiner, E. D., Baird, M. D., & Hamilton, L. S. (2015). Continued Progress: Promising Evidence on Personalized Learning. RAND Corporation.