“Education is the most powerful weapon which you can use to change the world”, Nelson Mandela.

Introduction

The far-reaching impacts of Artificial Intelligence in our lives, as we know them, are no longer a well-kept secret.¹ UNESCO’s AI Readiness Assessment notes that Africa’s Silicon Savannah, Kenya, has been at the forefront of incorporating AI in driving sectoral innovation and efficiencies across the agricultural, healthcare, education and financial services industries.² While countless success stories can be noted from the uptake of AI in Kenya, this piece narrows its focus down to the interaction of AI and moreso, the ethical impact of intelligent tutoring systems (ITS) in Kenya’s Primary Schools.

The term ITS refers to computer-based educational platforms designed to enhance the learning experience by providing tailored instructions that cater to individual student needs.³ These systems enhance student learning through standardised tests, assigning and grading homework as well as simulations to test a learner’s appreciation of given academic material.⁴ This piece’s central argument is that the continued implementation of ITS in Kenyan Primary Schools presents a unique opportunity to address the major policy concerns that arise, such as privacy considerations, algorithmic bias and the existing digital divide.

To better appreciate the importance of ITS in Kenya’s education sector, this piece employs a desk research methodology surveying relevant academic material to understand the general application of ITS. Based on this methodology, the findings of this piece are confined to Kenya’s primary schools, and the information relied on has been primarily drawn from anecdotal and secondary data. Having outlined these points, the piece continues by discussing the Kenyan situation with regard to ITS, the resultant ethical considerations and how the existing policy and regulatory framework addresses them. The piece then weaves these lessons into informing the recommendations intended to bridge the resultant gaps.

ITS in Kenya’s Primary School Education

Primary school students in Kenya have long been on the shorter side of the stick in terms of academic attention from their tutors. As recent as 2021, Kenya’s student-to-teacher ratio stood at 70:1 in some schools as compared to the recommended ratio of 40:1.⁵ This means that the average primary school teacher, in practically half of the counties in Kenya, teaches close to twice as many students as recommended by the Ministry of Education.⁶ A commonly agreed concern about this statistic is that these teachers fail to perform their professional duties effectively and that learners fail to receive the required academic attention at their level of study.

Given the prevailing nature of this challenge, several ITS have come to the fore, such as MwalimuPlus, Angaza Elimu, M-Shule and iMlango.

MwalimuPlus provides learner support for primary school students in Kenya by outlining theoretical overviews and practical examples.⁷ Angaza Elimu uses AI to provide value to both teachers and students. The MwalimuPlus platform allows teachers to identify a learner’s growth areas and assign bespoke learning resources, while the Angaza platform grants them access to resources and helps them to track their progress.⁸ In terms of access, Angaza Elimu’s ‘Kalamu’ initiative integrates ITS into the Competency-Based Curriculum (CBC) at affordable rates starting from Kshs 10 through a tripartite focus on adaptive and personalised learning and teacher intervention.⁹

M-Shule and iMlango are other commonly used in Kenya’s hard-to-reach rural areas, and the former is known to use an SMS-based system to facilitate academic delivery of lessons and evaluation tools to organisations.¹⁰ This system has navigated a common challenge of inadequate internet coverage. Another challenge to implementing ITS in Kenyan primary schools is insufficient teacher training. This challenge has made it difficult for current teachers to upload localised material for their learners, thus causing them to claw back on the intended impact of ITS.

To address this challenge, the Kenyan government, through the Ministry of Education, has embraced initiatives such as the Digital Skills Programme, led by the Information, Communication Technology Authority (ICTA), in 2022.¹¹ This programme is geared towards bolstering digital competencies, information literacy and data protection. So far, 80,000 public primary school teachers in Kenya have been trained in readiness for a technology-driven approach to learning.¹² While all the preceding may point towards the present successes of these systems, several ethical considerations arise from using ITS in Kenyan Primary Schools.

Ethical considerations and policy framework

Having said the foregoing, it behoves one to venture into the ethical concerns arising from the current subject. In the discourse surrounding artificial intelligence, the centrality of ethics cannot be overstated.¹³ The rapid evolution of AI technologies has only magnified the urgency of addressing ethical concerns, and ITS does not escape this predicament. These systems are often at odds with fundamental ethical principles such as transparency, privacy, accountability, and fairness, demanding rigorous scrutiny within the Kenyan context.

AI systems, including ITS, fundamentally challenge the traditional concept of individual data ownership.¹⁴ Through extensive data collection and data analysis, these systems accumulate detailed profiles of users, often surpassing the individual’s own awareness.¹⁵ This phenomenon – dubbed ‘epistemic privilege’ – raises profound privacy concerns, particularly in educational settings where instructors may access this data to tailor educational experiences.¹⁶ Such access, while intended to enhance learning outcomes, is exacerbated by the risk of privacy violations, with commonly constructed privacy policies offering little resolve due to their oft-complex nature.¹⁷

The issue of algorithmic bias within AI systems further complicates the ethical landscape.¹⁸ Bias, particularly within educational applications, poses significant challenges to fairness and equity.¹⁹ The deployment of ITS in educational institutions, designed to personalise learning experiences, may inadvertently perpetuate existing biases. This phenomenon is starkly illustrated by the 2020 GCSE grading scandal, where the use of AI in grading resulted in demonstrably unfair outcomes, disadvantaging students from less privileged backgrounds.²⁰ The risk of bias in ITS is especially pronounced in contexts like Kenya, where diverse student populations and varied cultural backgrounds heighten the potential for inequitable educational outcomes.

Transparency, a cornerstone of ethical AI deployment,²¹ remains elusive due to the inherent opacity of many AI systems.²² Often referred to as the “black box” problem, this lack of transparency undermines the trustworthiness of AI technologies.²³ In the context of ITS, transparency is indispensable for ensuring that the decisions made by these systems – decisions which have profound implications for a student’s educational trajectory – are comprehensible and justifiable. Without a clear understanding of how these systems operate and make decisions, stakeholders cannot fully trust or validate the outcomes they produce.

Compounding these ethical challenges is the pervasive digital divide, particularly pronounced in developing nations such as Kenya.²⁴ The digital divide, characterised by disparities in access to technology and digital infrastructure, poses a formidable barrier to the equitable implementation of ITS.²⁵ Infrastructural deficiencies, limited access to technological devices, and a lack of culturally relevant content exacerbate this divide, thereby impeding the broader adoption and effectiveness of AI-driven educational tools.²⁶

In addressing the myriad of ethical concerns, the role of policy and regulatory frameworks becomes paramount. Globally, frameworks such as UNESCO’s AI principles,²⁷ and the European Union’s General Data Protection Regulation (GDPR) emphasise the need for proportionality, fairness, privacy, and transparency in the deployment of AI technologies.²⁸ The recently instituted African AI framework also echoes these principles while concomitantly appreciating the importance of AI in education²⁹. Kenya’s Data Protection Act aligns with these global standards, providing a robust legal framework for regulating ITS. The Act mandates lawful, fair, and transparent data processing; also, explicit consent for data collection, and requires rigorous data protection measures.³⁰ However, the effectiveness of these regulations in addressing the unique risks posed by ITS within the educational sector remains an ongoing challenge.

Final remarks

Moving forward, this piece advances recommendations geared towards resolving the challenges canvassed above to realise a better state of ITS in Kenyan Primary Schools. Firstly, it is prudent to intensify investments into improving educational infrastructure, such as internet connectivity, through collaborative efforts of telecommunication companies and the government. The other recommendation geared towards the main drivers of this process, the teachers and students, is formulating bespoke training programmes for teachers and lessons for students. This fosters an environment of continuous learning from these vital parties.

Given the present wave of data protection, there should be a comprehensive review of the industry-specific data protection policies. This will ensure that data collection and processing methods are guided by solid principles, ultimately improving the quality of training data used to deliver this essential service. While education is indeed a powerful tool for spurring worldwide change, AI-driven Intelligent Tutoring Systems (ITS) enhance its effectiveness, making it even more accessible and driving innovation for the future.