Global Study of the Impact and Implications of Artificial Intelligence in Education in Developing Countries

Abstract

This paper explores the role of artificial intelligence in education (AIEd) in mitigating disparities in developing countries. Focusing on regions like India, Latin America, and Africa, the paper explores how AI technologies can address infrastructural limitations, such as limited internet access and language barriers, while also navigating cultural nuances¹’²’³’⁴ . Using a mixed-methods approach, combining literature review and case study analysis, this paper explores the following research questions: (1) How can AIEd be adapted to diverse cultural and infrastructural contexts in developing countries? (2) What are the potential impacts of AIEd on learning outcomes and teaching practices in these regions? Utilizing case studies, this paper demonstrates how AIEd initiatives, such as offline learning tools and personalized learning, improve student achievement rates and enhance teacher efficiency through automation of administrative tasks⁵.
The paper contrasts the application of AIEd in developed and developing nations, emphasizing the need for ethical frameworks, teacher training, and policy reforms to ensure sustainable implementation. It concludes that while AIEd offers transformative opportunities to improve accessibility and equity, its success depends on overcoming infrastructural and economic barriers through collaborative and context-specific approaches⁴’¹’³.

Keywords: Artificial intelligence in education (AIEd), developing countries, personalized learning, educational equity, case studies, mix-methods research

Introduction

Artificial intelligence (AI) has evolved significantly impacting various fields including education, evolving from basic automation to highly sophisticated systems capable of solving complex problems evolving from basic automation to highly sophisticated systems capable of solving complex problems¹. In the context of education, AI is increasingly recognized for its potential to personalize learning experiences and address diverse student needs around the world³’⁶.

This paper explores the transformative potential of AIEd in developing countries, focusing on regions like India, Latin America, and Africa. In this paper, the definition of developing countries describes countries often characterized by lower gross national income per capita, limited industrialization, and infrastructure compared to developed nations⁷. This paper investigates how AIEd can address pressing educational challenges in these contexts, aiming to answer the following research questions:

1. How can AIEd technologies be adapted to the unique cultural and infrastructural contexts of developing countries?
2. What are the potential impacts of AIEd on learning outcomes and teaching practices in developing countries, specifically addressing issues of education equity and access?

This study addresses a significant gap in current AIEd research by focusing on developing regions, where education systems often struggle due to economic and social limitations³. While existing research has explored AI in education broadly, there is a lack of comprehensive analysis that considers the specific challenges and opportunities presented by developing world contexts. While there have been a few studies looking at AIEd through a country or regional lens, this paper aims to fill a gap in the research by exploring AIEd initiatives and their impact globally across Africa, Latin America, and Asia. Additionally, this study seeks to bridge the divide between theoretical AI advancements and practical implications in resource-constrained educational settings such as the developing world. Finally, this research provides recommendations on how AI can effectively, ethically, and sustainably meet the unique educational needs of the developing world.

Employing a mixed-methods approach that combines literature review and case study analysis, this paper examines the specific challenges faced by developing countries, explores various AIEd technologies and their applications, and analyzes real-world case studies to provide evidence of their impact. Finally, the paper discusses the importance of tailored approaches and policy support, analyzes the implications of these findings, and offers recommendations for policymakers, educators, and researchers to ensure the effective and ethical integration of AI in education in the developing world.

Methodology

This research employed a mixed-methods approach to investigate the potential of AIEd to address educational challenges in developing countries. The methodology combined a systematic literature review to gather relevant academic articles, reports, and publications with a comparative case study analysis to provide a comprehensive understanding of AIEd implementation and its impact.

In terms of case study selection, case studies were selected based on their relevance to the research questions, geographic diversity, and representation of different AIEd technologies. Pilot projects and initiatives were chosen based on evidence of implementation and reported outcomes. Specific criteria for inclusion include: demonstrated implementation of AIEd technologies, availability of data on implementation and outcomes, and representation of diverse contexts within developing countries. Case study data were primarily collected through secondary sources, including published reports, evaluations, and website organizations involved in AIEd initiatives. To ensure the research remained focused on specific questions – namely, AIEd in developing countries as opposed to developed countries, recent AIEd developments, and to minimize selection bias, exclusion criteria were applied. 

Scope	Research on AIEd implementation in developing countries
Search Terms	“Artificial Intelligence in Education,” “AIEd,” “developing countries,” “personalized learning,” “intelligent tutoring systems,” specific country/region names (e.g., “AIEd India,” “AIEd Africa”), “case studies”
Databases Used	Science Direct, Web of Science, Google Scholar
Inclusion Criteria	Focusing on AIEd in developing countries, diverse contextsPublished after 2010Available in EnglishCountry or region-specific case studies on AIEdDemonstrated implementation of AIEd technologies Availability of data on implementation and outcomes
Exclusion Criteria	Focusing solely on developed countriesPublished earlier than 2010

A comparative case study analysis was conducted using an analytical framework that included dimensions such as technology type, target group, educational context, cultural adaptation, infrastructural challenges, outcomes implementation factors, and ethical considerations. Both quantitative and qualitative data were analyzed. Quantitative data, such as student achievement scores, were extracted from reports and analyzed to assess the impact of AIEd. Qualitative data such as descriptions of implementation processes and challenges were analyzed to provide context and insights. Where available, information on sample sizes and representativeness was included in the case study analysis.

Section 1: Understanding the Challenges in Education in the Developing World

The global education gap manifests in stark disparities across various indicators in the developing world. Access to education remains a significant challenge, with 243 million children and adolescents aged 6-18 excluded from schooling in 2021, predominantly in low and lower-middle-income countries⁸. Sub-Saharan Africa bears the brunt of this crisis, with over 98 million children out of school (UNESCO Institute for Statistics 2019). This lack of access is compounded by lower completion rates, particularly in primary education. In 2020, the average primary school completion rate in low-income countries was 66%, significantly trailing the 94% average observed in high-income countries, indicating a substantial proportion of children failing to attain even basic education⁹.

Exacerbating these challenges is a critical shortage of qualified teachers, particularly in remote and underserved areas, with pupil-teacher ratios far surpassing recommended levels for effective learning². This shortage is often coupled with inequitable resource distribution, with rural schools disproportionately lacking qualified teachers, adequate infrastructure, and essential learning materials¹⁰’⁹. These multifaceted challenges underscore the urgent need for interventions to bridge the global education gap and ensure equitable access to quality education for all.

Resource Scarcity and Impact

One of the key challenges faced by developing countries is resource scarcity³’¹. This comprises financial constraints and a lack of essential resources such as qualified teachers, learning materials, and adequate infrastructure². Resource-constrained environments create a challenging landscape for effective learning³. Overcrowded classrooms and limited access to textbooks and technology can severely hinder the learning environment and negatively affect student achievement²’⁴. Studies have shown a direct correlation between resource availability and student performance, with students in under-resourced schools consistently lagging behind their peers with adequate resources (Glewwe & Muralidharan 2016).

To address these challenges, various strategies have been implemented in developing countries. These include initiatives to improve teacher training and student/teacher ratios and financial assistance to disadvantaged students and schools²’¹. However, these efforts often face limitations due to the scale and lack of financial resources¹¹’¹.

By leveraging technology to personalize learning experiences, providing access to quality education in remote areas, and supporting teachers with data-driven insights, AIEd can offer innovative solutions to address the challenges posed by resource scarcity (Chisom et al. 2023). For instance, AI-powered learning platforms can provide personalized learning experiences even in overcrowded classrooms, while offline AIEd tools can bridge the gap in areas with limited internet connectivity (Hakimi & Shahidzay 2024; Chisom et al. 2023). Furthermore, AIEd can assist in optimizing resource allocation, ensuring that limited resources are used effectively to maximize their impact²’¹².

While these challenges are widespread, developing countries face unique obstacles that require targeted solutions that are different in developed country contexts. In India, educational inequity is particularly pronounced. Children from lower castes, those in rural areas, and girls face significant disadvantages due to deep-rooted social hierarchies¹³. According to a public report on basic education in India:

“While India has made significant strides in expanding access to education, the quality of education, especially in government schools, remains a serious concern. This is reflected in low learning outcomes, high dropout rates, and a lack of preparedness for the demands of the 21st-century workforce.”¹⁴.

The education sector in India, comprising about 340 million students, has a massive demand and opportunity to transform itself and cater to future work demands⁶.

In Latin America, while access to education has improved, poverty and marginalization continue to affect access to quality education, particularly for indigenous populations, children with disabilities, and those living in rural areas¹⁵. Lustig et al. (2016) state that poverty creates a vicious cycle, with children from low-income families being less likely to attend preschool. This creates a cumulative disadvantage, leading to early dropout and limited future opportunities. This was also validated by Salas-Pilco & Yang (2022), who studied higher education and concluded that students from lower-income households are more likely to discontinue their studies. While technology can improve learning outcomes, Trucco & Palma (2021) highlight the digital divide in Latin America, especially in rural and low-income areas, where many schools lack adequate infrastructure and connectivity.

Africa, with its rich diversity of cultures, languages, and socio-economic conditions, faces educational challenges such as poverty and child labor, linguistic diversity, and resource constraints. Despite progress in enrollment, poverty remains a major barrier. Many children, particularly in sub-Saharan Africa, are forced into child labor to support their families, hindering their access to education¹⁶. Navigating linguistic diversity is also a challenge. Many African countries face difficulties implementing multilingual education policies due to limited resources, inadequate teacher training, and the dominance of colonial languages in education systems². Studies have highlighted the importance of educating students in their mother tongue in early grades for improved learning outcomes and have shown that language barriers can hinder access to education for minority language groups¹⁷. Additionally, children displaced by war and living in refugee camps, particularly in conflict zones, face limited access to education. These children often rely on innovative solutions like radio and AI-powered learning to continue their education despite resource shortages and unsafe environments (Global Partnership 2023). Finally, overcrowded classrooms, a lack of learning materials, and inadequate infrastructure are resource constraints that hinder the quality of education in these countries (Glewwe & Muralidharan 2016).

Overcoming these challenges is crucial for achieving sustainable development goals. The next section will explore AI in education in the developing world and how it offers the potential to address persistent educational challenges such as limited access, scarce resources, and diverse learning needs.

Section 2: AIEd Technologies and Their Applications

AIEd is gaining momentum, particularly in the developing world, offering the potential to transform education and accelerate progress towards global education goals¹⁸. There is a strong belief that AI can profoundly change the education sector and accelerate the achievement of global education goals by reducing barriers to accessing learning, automating management processes, and optimizing methods to improve learning outcomes¹⁸. This section examines prevalent AIEd technologies/applications and their specific applications in addressing the challenges discussed in Section 1 while acknowledging their limitations in the context of developing countries. This article focuses on personalized learning systems, as they are widely discussed and encompass a broader category of AI applications/technologies such as intelligent tutoring systems and chatbots.

Personalized Learning Systems (PLS)

Personalized learning systems (PLS), also known as Personalized Learning Environments (PLE), are software platforms, running on personal computers, tablets, or phones, that use AI algorithms to analyze large amounts of data, including students’ progress, performance, and preferences to tailor educational content, pacing, and instructional strategies to individual students’ needs, preferences, and learning styles¹⁹.

By adapting content to each learner’s unique needs, PLS aims to create a more engaging, effective, and efficient learning process. These systems typically incorporate features such as adaptive instruction, feedback, data analytics, and positive reinforcement¹⁹. Adaptive systems ensure an individualized learning experience unique to each learner while providing immediate feedback, enabling students to identify and rectify errors promptly, and creating a cycle of continuous learning and self-improvement. Integral to these systems is the utilization of data analytics. By capturing and analyzing learning behaviors, these systems provide educators with valuable insights into student progress and areas for potential intervention. This data-driven approach informs instructional decisions, curriculum design, and assessment strategies, promoting a more responsive and effective learning environment. Finally, adaptive learning systems emphasize student motivation and engagement through the integration of positive reinforcement and game-like elements. This gamified approach, coupled with personalized content and activities, aims to foster a more enjoyable and stimulating learning experience.

This technology directly addresses the challenge of resource scarcity by providing individualized learning paths even in overcrowded classrooms with limited teacher availability²’⁴. PLS can also help overcome inequitable distribution of resources by providing access to quality learning materials regardless of student’s location or socioeconomic background¹⁹’⁴. However, limitations such as the need for reliable internet connectivity pose challenges to PLS implementation in developing countries¹’²⁰.

Intelligent Tutoring System (ITS)

Intelligent tutoring systems (ITS) represent a specialized subset of PLS designed to provide personalized instruction, guidance, and feedback similar to a human tutor, creating a responsive learning environment that adapts to individual needs²¹. These systems can support a variety of tasks, including cognitive skills development for students working on problem-solving skills, writing tutors that provide feedback on grammar, style, and organization in written work, and language learning tutors that offer personalized instruction and practice in foreign languages. Intelligent tutors hold significant promise for transforming education by providing personalized, adaptive, and engaging learning experiences²¹. This can be particularly beneficial in developing countries where there is a shortage of qualified teachers²²’⁶. ITS can also help address linguistic diversity by offering instruction in multiple languages²’⁴. However, challenges include the cost of development and deployment, and the need for culturally relevant content to ensure effective learning³.

Chatbots

Chatbots are computer programs designed to mimic human dialogue and utilize natural language processing (NLP) to interact with students and provide support²³’²⁴. Machine learning (ML) algorithms enable chatbots to learn from past interactions and personalize the learning experience (Hobert & Cocea 2017). While their use in education is still limited, chatbots offer the potential for personalized support, increased engagement, 24/7 availability, and scalability²³’²⁵.

This can be valuable in addressing the challenge of limited access to education, as chatbots can provide 24/7 support and guidance to students even outside of school hours. However, limitations include the complexity of developing sophisticated chatbots that can effectively understand and respond to diverse student needs and concerns about data privacy and security.

Title	Author	Year	AI – Tech nology	Educa tional Challenges	Developing Countries	AI Technology Focus	Educational Challenges Addressed	Country/ Region Focus
Review of AI in Education: Transforming Learning Environments in Africa	Chisom, O. et al.	2023	X	X	X	Personalized learning system, Chatbots, Intelligent Tutoring System	Disparities in access, quality, and infrastructure, language barriers, shortage of qualified teachers, and high drop out rates	Global with various countries/ regional examples
Potential of Artificial Intelligence for transformation of the education system in India. The International Journal of Education and Development using Information and Communication Technology, 17 (1), 142-158.	Jaiswal, A. & Arun, J.	2021	X	X	X	Personalized learning and recommendation system, Adaptive assessment	Inequity, low access, and poor quality of education	India
Personalized education and Artificial Intelligence in the United States, China, and India: A systematic review using a Human-In-The-Loop model	Aditi Bhutoria	2022	X	X	X	Personalized learning system, Chatbots, Intelligent Tutoring System	Need for personalized learning (vs one size fits all), and teacher overload	US, China and India
Artificial intelligence applications in Latin American higher education: a systematic review	Salas – Pilco & Yang	2022	X	X	X	ChatbotsAI computer -assisted content analysisPredictive modelingImage Analytics	Educational inequality, and socio-economic disparities lead to dropouts	Latin America
AI technologies for education: Recent research & future directions. Computers and Education: Artificial Intelligence, 2, 1-11.	Zhang, K., & Aslan, A	2021	X	X	X	Personalized learning systemChatbotsIntelligent Tutoring SystemExpert systemsMachine LearningVisualizations	Significant gap between AI technological innovations and educational applications, lack of educational perspectives in AIED research, costs & scalability, ethical and privacy concerns, need for more empirical research, limited research coverage of specific AI technologies, lack of actionable guidelines and AI expertise for educators	Global with various country/ regional examples
Transforming Education with Artificial Intelligence: Potential and Obstacles in Developing Countries	Musawer Hakimi & Amir Kror Shahidzay	2024	X	X	X		Technological barriers such as limited access to internet, infrastructural challenges, socio-cultural factors, ethical and data concerns, and challenges of meeting individual needs	Global with various country/ regional examples
Assessment of Students’ Experiences and Viewpoints in Using Chatbots for Education Practice: A University Case of a Developing Country	Moyo, S. et al.	2024	X	X	X	Chatbots	High rate of lecturer turnover, need for clearer policies and guidelines for AI adoption, need for ethical considerations, addressing negative perceptions and privacy concerns	Zimbabwe
AI in Education: A Systematic Literature Review	Tahiru, Fati	2021	X	X		AutomationSmart ContentIntelligent Tutoring Systems	Technological challenges related to the cost and maintenance of AIED systems, data privacy and security Concerns, Organizational readiness, lack of research in developing countries, need for more empirical studies
AI in education: A review of personalized learning and educational technology. GSC Advanced Research and Reviews, 18(02), 261–27. https://doi.org/10.30574/gscarr.2024.18.2.0062	Ayeni, O., Al Hamad, N.M., Chisom, O.N., Osawaru, B.	2024	X	X		Personalized Learning System	Privacy concerns, digital divide and accessibility issues, and ethical guidelines
Digital Education Revolution: Navigating the Challenges of AI Integration in Developing Nations	Robert Thomson & Julia Anderson	2023	X	X	X		Limited infrastructure, digital divide, access and equity concerns, ethical considerations, scarcity of experience in AIEd
Artificial Intelligence in Developing Countries: Bridging the Gap Between Potential and Implementation	Aderibigbe, A. et al.	2023	X	X	X		Infrastructural and digital divide, lack of skills in AIEd, inadequate educational curricula, ethical and societal considerations, underlying resource constraints in providing quality education
Artificial intelligence in developing countries: The impact of generative artificial intelligence (AI) technologies for development	Mannuru, N.R. et al.	2023	X	X	X	Chat GPTDALL – E	Infrastructure and technological disparity in developing countries, lack of skills in AIEd, ethical and societal considerations, and robust regulatory frameworks

Section 3: AIEd in the Developing World: Case Studies and Evidence

While AIEd technology application is still in its infancy, this section examines case studies of AIEd implementation in India, Latin America the Caribbean, and Africa, analyzing their approaches, limitations, and potential biases. It also provides a more in-depth analysis of the impact of AIEd on learning outcomes and teaching practices, incorporating quantitative data where available.

India

AIEd is being explored in India to address challenges related to educational inequity, quality, and access, particularly in rural and under-resourced communities⁶. Zaya Learning Labs, a non-governmental organization (NGO), for instance, uses AI to deliver personalized learning experiences focused on literacy and numeracy skills, with promising results in pilot projects conducted in underserved areas²⁶. However, potential drawbacks include the scalability of the program and reliance on and access to technology.

Another company, Educational Initiatives, uses an AI software called Ei Asset to help evaluate a student’s basic understanding of concepts by asking unfamiliar and thought-provoking questions, preventing rote learning and encouraging students to understand subjects in depth²⁷. According to the website, 15 million assessments have been conducted using EiAsset²⁷. Another tool Ei Mindspark, provides adaptive and personalized learning. A randomized control trial of Ei Mindspark by J-Pal shows 2-3x gains in students’ learning outcomes relative to control schools and a 7-8x gain in DIB Lucknow (https://ei.study/ei-mindspark). While learning gains are significant, limitations for EiMindspark include the necessity of consistent internet access for optimal functionality. Additionally, ensuring equitable access to devices for all students to utilize the platform effectively is another hurdle.

Figure 1: Randomized control trial of Ei Mindspark by J-Pal shows 2-3x gains in students’ learning outcomes relative to control schools and a 7-8x gain in DIB Lucknow (https://ei.study/ei-mindspark/)

The company has other AI projects including Ei Reading a similar tool, that measures reading literacy levels and diagnoses their strengths and weaknesses in essential reading skills and Ei CARES (an assessment, report, and evaluation tool for students to help replace conventional assessments and free up time for teachers)²⁸. However, limitations may include the need for teacher training to effectively utilize the data generated by the software⁶.

Latin American and the Caribbean

Initiatives like Oppia and Kolibri offer examples of how AIEd can be leveraged to provide accessible and personalized learning opportunities. Oppia, based in Brazil, is a multilingual and personalized platform offering free education to underprivileged learners (Myers et al. 2022). However, Oppia also has drawbacks due to its reliance on Internet connectivity. Kolibri has addressed this in its open-source platform by providing content offline, catering to learners in low-resource settings with limited internet connectivity (Myers et al. 2022)². Both the free access and offline functionality are critical in bridging the digital divide and ensuring educational access in remote areas. Chisom et al. (2023) note that “pilot projects in various countries, especially those with limited internet access, have demonstrated the effectiveness of AI in delivering educational content offline”².

Africa

A notable case study in South Africa involved the implementation of the AI-powered personalized learning system Mindspark in under-resourced primary schools²⁹. Mindspark uses AI algorithms to create personalized learning paths in Math and English, provides teachers with data on student performance, and offers accessibility for both in-school and at-home learning. Data showed that Mindspark was able to demonstrably improve student understanding and academic progress (with student scores showing significant improvement in both Grade 5 and Grade 6 with average increases of 6.5% and 10.8% respectively compared to students who received traditional instruction)²⁹. Another positive outcome was increased engagement with over 60% of students accessing the platform outside of school hours²⁹.

However, it is important to acknowledge that the findings from Mindspark implementation in South Africa cannot be generalized to all African countries or the broader developing world. The vast differences in infrastructure, cultural contexts, and education system across these regions warrant some caution when extrapolating the results. While Mindspark has shown promise in South Africa (and is being expanded to Botswana and Zambia), further research is needed to evaluate the impact in diverse contexts across Africa and beyond.

As Chisom et al. (2023, p. 649) argue, “The overarching theme of personalization emerges as a central tenet, as AI-driven adaptive learning platforms cater to individual student needs, addressing the diverse learning profiles present across Africa.” This personalized approach, coupled with real-time feedback and teacher support, promotes a more responsive and student-centered learning environment.

3.5 Impact on Learning Outcomes

Studies have shown that AIEd can significantly improve learning outcomes and accelerate learning. In India, quantitative data shows students using an AI-powered system ITS completed the curriculum in half the time compared to traditional methods, with comparable learning outcomes⁶; Shute, Glaser, & Raghavan 1998). In South Africa, Mindspark increased the test scores of students by 6.5% and 10.8% compared to traditional instruction. In Latin America, AI facilitators are used to provide real-time support to teachers in the classroom¹². Adaptive learning systems like ITS not only personalize content but allow for varied learning paces, and support diverse learning needs and styles which is especially valuable in classrooms with large student populations, and limited resources³. However, more studies are needed on the quantitative and qualitative impact as well as long-term results.

3.6 Impact on Teaching Practices

AIEd technologies present a significant opportunity to revolutionize teaching practices in developing countries. By streamlining administrative processes like grading and reporting, AI can free up educator’s time for more impactful activities such as providing individualized support and enhancing teaching strategies²’¹²’⁶. These systems analyze diverse student data, offering valuable insights into individual learning styles, needs, and potential challenges. This capacity enables teachers to tailor instruction, implement timely interventions, and enable early identification of at-risk students holds promise to improve retention and graduation rates¹².

The integration of AIEd offers a powerful view into student learning, providing data-driven insights that allow teachers to understand individual learning trajectories and address specific gaps²³’¹⁹’².  While beneficial, the potential for teacher displacement due to automation is a concern, although, in areas with limited teacher resources, it could also be seen as an advantage³⁰. A more significant challenge in the developing world is the high cost of implementing and maintaining these technologies. Effective adoption depends on adequate teacher training to utilize AI tools, interpret data, and integrate them into pedagogy⁶.

AIEd technologies can also optimize resource allocation by assisting with administrative tasks such as scheduling and resource deployment, leading to more efficient use of teaching staff. Salas-Pilco & Yang (2022) highlight the diverse applications of AIEd, including the use of chatbots to gather student feedback on teaching practices, fostering a culture of continuous improvement within educational institutions.

In essence, AIEd’s ability to automate tasks and provide data-driven insights empowers teachers to focus on individual student needs and refine their teaching (Chisom et al. 2023). While still in the early stages, these applications in developing countries show promise for enhancing learning outcomes, student engagement, and accessibility by improving learning experiences and offering opportunities for task automation and timely interventions.

While still at a nascent stage, emerging applications of AIEd in developing countries, such as personalized learning systems and intelligent tutors, demonstrate early promise in enhancing learning outcomes, student engagement, and accessibility. These technologies offer significant advantages for both students and teachers, including improved learning experiences and opportunities for educators to automate tasks and provide timely interventions.

Table 3 provides an analytical framework for comparing case study examples across various areas including technology type, cultural adaptation, infrastructure challenges, outcomes and ethical considerations.

Dimension	Case Study 1: Zaya Learning Labs (India)	Case Study 2: Oppia (Brazil)	Case Study 3: Mindspark (South Africa)
Technology Type	Personalized learning system	Personalized platform, Multi-lingual	Personalized learning system
Target Group	Primary students in under-resourced areas	Underprivileged learners	Primary students in under-resourced schools
Educational Context	Rural and under-resourced areas	Low-resource schools	Low-resource schools
Cultural Adaptation	Tailored to foundational education –  literacy and numeracy skills	Multilingual platform	Personalized learning paths in Math and English
Infrastructural Challenges	Limited access to technology and internet connectivity	Limited internet connectivity	Limited access to technology and internet connectivity
Outcomes	Improved learning outcomes in pilot projects	Increased access to education	Improved student understanding and academic progress; increased engagement
Ethical Considerations	Equitable access	N/A	N/A

Result Summary

This section summarizes the key findings from the reviewed literature and the case study analysis of AIEd in developing countries. The evidence suggests a consistent trend across India, Latin America and the Caribbean, and Africa indicating the potential of AIEd to address critical education challenges.

Firstly, personalized learning systems (PLS) and intelligent tutoring systems (ITS) have demonstrated promise in improving learning outcomes as evidenced by the Ei Mindspark study in India showing 2-3x gains (https://ei.study/ei-mindspark), ITS study in India showing curriculum completion in half the time⁶, and Mindspark South Africa results showing increased the test scores of students by 6.5% and 10.8% compared to traditional instruction²⁹. These gains are attributed to the ability of AI to tailor content and pacing to individual student needs, providing immediate feedback and support, which is particularly valuable in large and under-resourced classrooms²⁰’².

Secondly, AIEd technologies such as chatbots and automated assessment tools like Ei Asset and Ei CARES, are shown to enhance teacher efficiency by automating administrative tasks. This allows educators to dedicate more time to direct student interaction and teaching improvements²’¹². However, the effectiveness of these tools is contingent on adequate teacher training and seamless integration into existing teaching practices.

Thirdly, the case studies highlight the importance of adapting AIEd technology solutions to the specific infrastructural and cultural contexts of developing countries. The success of offline platforms like Kolibri in regions with limited internet connectivity and the focus on multilingual content in initiatives like Oppia underscore the need for context-aware design and implementation. The varying cultural values and socio-economic conditions across these regions require careful consideration to ensure social acceptance and equitable access to AIEd technologies.

Finally, while the potential benefits are evident, the reviewed research also identifies significant limitations. These include the digital divide, the high cost of implementation and maintenance, the need for robust infrastructure, and ethical concerns related to data privacy and algorithmic bias. Addressing these challenges is crucial for the sustainable and equitable scaling of AIEd in developing countries.

Discussion

Artificial intelligence in education (AIEd) is reshaping the educational landscape in terms of how education is delivered and accessed globally. As highlighted in the case studies, AIEd in developing countries is uniquely positioned to tackle foundational challenges related to access, equity, and resource distribution contrasting with the focus on enhancement and efficiency often seen in developing countries³⁰’³.

The quantitative data from the case studies provide concrete examples of AIEd’s impact. For instance, the J-Pal study on EI Mindspark in India demonstrated a significant increase (2-3x) in learning outcomes in control schools and an even more substantial 7-8x gain in a specific location (DIB Lucknow). Similarly, the implementation of Mindspark in South African primary schools resulted in average test score increases of 6.5% in Grade 5 and 10.8% in Grade 6²⁹. The study, cited by Jaiswal and Arun (2021) indicated that students using an AI-powered ITS in India completed the curriculum in half the time with comparable learning outcomes. These figures underscore the potential of AIEd to accelerate learning and improve academic performance in resource-constrained environments.

Cultural and Socio-Economic Factors

The effectiveness of AIEd is significantly influenced by cultural and socio-economic factors. In culturally diverse regions like Africa, Latin America, and India, the need for AIED tools that are linguistically appropriate and culturally sensitive is paramount². For example, implementing a learning platform that only operates in English in a region with numerous local languages will limit its effectiveness. Social-economic disparities also play a crucial role. The digital divide, as highlighted in the context of Latin America³¹, means that students from lower-income backgrounds may lack access to the necessary devices and internet connectivity to fully benefit from online AIEd resources. Even offline solutions like Kolibri need accessible devices, which can be a barrier for the most marginalized populations.

Consider the example of gender disparities in technology access in some developing countries. If girls have less access to smartphones or tablets due to cultural norms or economic constraints, AIEd interventions through these devices may inadvertently exacerbate existing gender inequalities in education¹². Furthermore, the cultural perception of technology and the role of teachers can influence the adoption and effectiveness of AIEd requiring careful community engagement and teacher buy-in. In some communities, social acceptance continues to be a challenge, with factors such as gender gaps and concerns about the changing role of teachers influencing AIEd implementation rates¹²’².

Considering the economic significance of AIEd, while it can reduce educational costs over time by decreasing dependency on physical resources and facilitating online learning, the initial investment is often difficult in lower-income regions. Affordability of technology and sustainable infrastructure are important, as they determine the extent to which AIEd can be effectively implemented in areas where access to electricity and the internet is limited¹’³²’³. As Aderibigbe et al. (2023) highlight:

“To bridge the gap, significant investments in digital infrastructure are crucial. Governments and private entities should collaborate to improve broadband connectivity, ensure reliable power supply, and establish data centers. Accessible and robust infrastructure forms the foundation of effective AI implementation” (p. 192).

Impact of AIED on Students

Beyond test scores, AIEd can impact students in various ways. Personalized learning systems can increase student engagement and motivation by providing content that is relevant to their interests and learning styles¹⁹. The immediate feedback offered by ITS can foster a growth mindset and reduce learning anxiety. For instance, a student struggling with a math concept can receive targeted support and practice toward learning. AIEd tools can cater to diverse learning needs, supporting students with disabilities or different learning paces, and promoting a more inclusive education environment.

However, it is crucial to also consider potential negative impacts. Over-reliance on AIEd without sufficient human interaction could affect students’ social and emotional development. Concerns about data privacy and the ethical use of student data collected by AI must also be carefully addressed to ensure student well-being and trust in these technologies.

Finally, the application of AIEd raises ethical questions, particularly regarding privacy, potential biases in AI algorithms, and the preservation of cultural values³³’¹’²’³. These concerns are especially relevant in countries where frameworks around AI are less developed, making it essential to establish ethical guidelines to protect students’ data and privacy.

Recommendations for the Developing World

To maximize the impact of AIEd, some recommendations are proposed to address specific concerns in developing countries. These include adapting content to local contexts and linguistic needs, and ensuring the appropriate infrastructure is in place to successfully leverage these technologies.

Firstly, localized AI development is essential to ensure that AIEd technologies meet specific educational needs and adapt to the different cultural contexts of developing regions¹. AIEd technologies used in developing countries shouldn’t simply replicate existing systems from other regions; instead, they should be tailored to address the specific educational, cultural, and linguistic needs of local communities¹. For example, integrating culturally relevant examples into AI learning systems can increase student engagement and make educational content more relatable. Thompson and Anderson (2023) emphasize this point, stating that AIEd requires “an approach that is not only innovative but also grounded in the practical realities of each unique educational landscape” (p. 6). In India, AI is being used to overcome language barriers in education by promoting the translation of textbooks into Indian languages using AI tools like ChatGPT. This is important because many students struggle to understand complex subjects in English, which can lead to exclusion. This approach aims to create a more inclusive education system where students can learn in their mother tongue. This is not only beneficial for students but also for the country’s overall development. Governments, NGOs, and technology companies must collaborate to create relevant AI-driven content that resonates with diverse populations and educational levels. Encouraging teachers to participate in the content creation process can also ensure that these resources resonate with the students they are designed for. By focusing on these targeted initiatives, AIEd can better address the unique educational challenges of developing countries, creating a more inclusive and sustainable approach to AI in education.

Investment in digital infrastructure is another critical factor for successful AIEd implementation. As Aderibigbe et al. (2023) emphasize, governments in developing regions must prioritize funding to provide reliable internet access, affordable devices, and electricity in underprivileged areas, creating a stronger foundation for AI-based education. Furthermore, to reach students in areas with limited internet access, AI education technology must be available offline³⁴’³²’³. The case studies in Latin America illustrate how offline AIEd resources, such as ITS, allow students in remote areas to benefit from AI-enhanced learning experiences¹². The availability of such tools is crucial for providing equitable educational opportunities and meaningful learning experiences in underprivileged areas. According to Mannuru et al. (2023): “Studies in developing countries, who already face limited resources and educational barriers, may encounter further marginalization if they lack access to the necessary technology, infrastructure or training required to leverage Generative AI effectively, which could deepen the digital divide and perpetuate existing inequalities in education.” (p. 7).  By developing this infrastructure, developing countries can overcome significant barriers to implementing AIEd technologies effectively.

The Role of Teachers

Teachers play a vital role in AIEd implementation, especially in developing regions where educational gaps exist. While AIEd technologies offer support by automating tasks and providing valuable insights, teachers still face challenges in integrating these tools into the classroom³. According to Hakimi & Shahidzay (2022) “there is a need for substantial AI instruction to be given to teachers in order to develop capacity as well as confidence in effectively using AI tools in the classroom” (p 9). AIEd requires teachers to adopt new technologies and adapt their teaching methods, which can be demanding. Teachers’ adaptability is therefore crucial for the success of AIEd initiatives in developing regions.

To prepare teachers for this shift, comprehensive professional development programs are necessary. These programs should focus on equipping teachers with the skills and strategies required to use AI tools effectively. Training that emphasizes the practical applications of AIEd in lesson planning and instruction can empower teachers to leverage these technologies to their fullest potential. By building the skills to use AIEd resources confidently, teachers can provide more engaging and personalized learning experiences for their students.

As AIEd begins to automate certain tasks, teachers’ roles may evolve from direct instruction to more mentoring and facilitative roles²³. This change emphasizes the importance of teachers as mentors, offering human support alongside the capabilities of AI technologies. By focusing on individualized student support and guidance, teachers can contribute the crucial human element to education that AIEd cannot replace. Therefore, AIEd should be viewed as a resource that enhances, rather than replaces, the role of teachers in students’ academic and personal growth.

The Role of Policy Advisors

Policy advisors also play a critical role in evaluating the effectiveness and ethical implications of AIEd on a broader level. They are crucial in assessing AIEd’s impact on local communities and advancing research to understand its long-term effects. These advisors ensure that AIEd initiatives align with the educational needs and values of the communities they serve. By leading research efforts and encouraging policy reforms that address implementation challenges, policy advisors can support the sustainable integration of AIEd in developing countries.

To fully assess AIEd’s effectiveness, policy advisors should prioritize local research initiatives that gather direct feedback from students, teachers, and parents. This approach allows for a deeper understanding of AIEd’s impact, ensuring that policymakers can continually improve AI tools and initiatives, fostering a more inclusive and adaptable educational system. Further studies are also needed to assess the effects of AIEd on learning engagement and long-term educational outcomes.

Limitations and Avenues for Future Research

Despite the valuable insights offered in this research, several limitations should be acknowledged. Firstly, this study relied heavily on secondary data sources for case study analysis. This reliance restricts the depth of insights and may introduce potential biases due to varied organizational reporting practices and potential data privacy concerns. Future research should prioritize the collection of primary data, including longitudinal studies, to provide a more comprehensive understanding of the long-term impacts of AIEd interventions.

Secondly, the availability of comprehensive academic research on AIEd in developing countries is limited. This necessitated the inclusion of some non-academic sources, which, while providing valuable real-world applications, may not always meet rigorous academic standards. Future research could address this limitation by conducting data-driven research using quantitative and qualitative methods to gather more reliable, evidence-based data on its effectiveness in various educational contexts.

Thirdly, detailed information on sample sizes for programs like Zaya Learning Labs and Mindspark was not consistently available, limiting the ability to assess the statistical significance and generalizability of the findings. Researchers should strive to gather and report detailed demographic and sample size data to enhance the rigor of their findings.

Future research should address the limitations identified in this study and delve deeper into the nuances of AIEd implementation in developing countries. First, longitudinal studies on learning outcomes that track the academic progress and long-term attainment of students using AIEd interventions compared to control groups would provide more robust evidence of the sustained impact of these technologies. Second, conduct in-depth qualitative research on student and teacher experiences using qualitative methods such as in-depth interviews and focus groups to explore perspectives of students and teachers using AIEd, their engagement levels, and educational outcomes. This will provide rich insights into the perceived benefits, challenges, and socio-emotional impacts of these technologies. Third, conduct more localized studies that examine the effectiveness of AIEdt tools within particular cultural and socio-economic contexts. This research should explore how cultural values, language, and access to resources mediate the impact of AIEd. Fourth, future research should compare the effectiveness of different AIED modalities (e.g. PLS, ITS, chatbots) in various education settings within developing countries to identity which technologies are most suitable for specific contexts and learning objectives. Fifth, conduct rigorous studies to assess the impact of AIEd on educational equity and access for marginalized groups, including girls, students from low-income backgrounds, and those in remote areas. This research should identify potential disparities in access and outcomes and propose strategies to mitigate them. Sixth, given the infrastructural challenges in many developing countries, more research is needed on the design implementation and effectiveness of offline AIEd solutions and strategies for their sustainable development. Finally, future studies should focus on developing standardized evaluation frameworks for AIEd initiatives, including both quantitative and qualitative measures to facilitate cross-case comparisons and meta-analyses.

Specific recommendations for policymakers, educators, and researchers are found in Table 4 below.

Stakeholder Group	Recommendations
Policy Makers	Invest in robust digital infrastructure, including reliable internet connectivity and access to devices, to create an enabling environment for AIEd. Develop and implement ethical guidelines for AIEd, addressing issues of data privacy, algorithmic bias, and cultural sensitivity. Promote public-private partnerships to support AIEd initiatives and ensure sustainable funding.
Educators	Participate in professional development programs to enhance their skills in using AIEd technologies effectively. Collaborate with researchers and technology developers to adapt AIEd tools to the specific needs of their students and communities. Advocate for the integration of AIEd into teacher training curricula.
Researchers	Conduct rigorous evaluations of AIEd interventions, including randomized controlled trials and longitudinal studies. Develop and validate culturally relevant AIEd tools and content. Investigate the ethical implications of AIEd and develop strategies to mitigate potential risks. Focus on gathering long-term data, and creating standardized evaluation metrics.

Conclusion

AIEd technologies have the potential to address common challenges such as access to education, equity, and resource distribution, particularly in regions facing inadequate infrastructure and limited educational opportunities. This study highlights the importance of context-specific solutions to overcome these challenges and emphasizes the important role that localized AI development, teacher adaptation, and strong policy frameworks play in ensuring the successful integration of AIEd. By proposing targeted recommendations, including investments in digital infrastructure and offline compatibility, this paper contributes to the ongoing discussion of how AIEd can be leveraged to improve educational outcomes in developing countries. Furthermore, it emphasizes the need for collaborative efforts between governments, NGOs, and other companies to create a sustainable and inclusive approach to AI implementation in education. Ultimately, this study provides a comprehensive understanding of the challenges and opportunities of AIEd in developing regions, contributing valuable insights for policymakers, educators, and researchers seeking to use AI to improve global education equality.

References

1. Aderibigbe, A., Ohenhen, P., Nwaobia, N., Gidiagba, J., Ani, E. (2023). Artificial Intelligence in Developing Countries: Bridging the Gap between Potential and Implementation. Computer Science & IT Research Journal, 4(3), 185-199. https://www.researchgate.net/publication/376196753_ARTIFICIAL_INTELLIGENCE_IN_DEVELOPING_COUNTRIES_BRIDGING_THE_GAP_BETWEEN_POTENTIAL_AND_IMPLEMENTATION [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
2. Chisom, O., Unachukwu, C., & Osawaru, B. (2023). Review of AI in Education: Transforming Learning Environments in Africa. International Journal of Applied Research in Social Sciences, 5 (10), 637-654. https://fepbl.com/index.php/ijarss/article/view/725 [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
3. Hakimi, M. & Shahidzay, A. K. (2024) Transforming Education with Artificial Intelligence: Potential and Obstacles in Developing Countries. Preprints , 2024072542. https://www.researchgate.net/publication/383129232_Transforming_Education_with_Artificial_Intelligence_Potential_and_Obstacles_in_Developing_Countries [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
4. Mannuru, N. R., Shahriar, S., Teel, Z. A., Wang, T., Lund, B. D., Tijani, S., Pohboon, C. O., Agbaji, D., Alhassan, J., Galley, J., Kousari, R., Ogbadu-Oladapo, L., Saurav, S. K., Srivastava, A., Tummuru, S. P., Uppala, S., & Vaidya, P. (2023). Artificial intelligence in developing countries: The impact of generative artificial intelligence (AI) technologies for development. Information Development, 0(0). https://doi.org/10.1177/02666669231200628 [↩] [↩] [↩] [↩] [↩] [↩]
5. Jet Education Services 2024; Chisom et al. 2023; Salas-Pilco & Yang 2022; Jaiswal & Arun 2021 [↩]
6. Jaiswal, A. & Arun, J. (2021). Potential of Artificial Intelligence for transformation of the education system in India. The International Journal of Education and Development using Information and Communication Technology, 17 (1), 142-158. https://www.researchgate.net/publication/349172956_Potential_of_Artificial_Intelligence_for_transformation_of_the_education_system_in_India [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
7. World Trade Organization. (2023).Who are the developing countries in the WTO? https://www.wto.org/english/tratop_e/devel_e/d1who_e.htm [↩]
8. UNESCO. (2022). 244 million children won’t start the new school year. Global Education Monitoring Report. https://www.unesco.org/gem-report/en/articles/244m-children-wont-start-new-school-year#:~:text=Paris%2C%201%20September%202022%20%E2%80%93%20As,are%20still%20out%20of%20school [↩]
9. World Bank. (2018). World development report 2018: Learning to realize education’s promise. World Bank Publications [↩] [↩]
10. UNICEF. (2016). Fixing the broken promise of education for all: Findings from the global initiative on out-of-school children. UNICEF [↩]
11. Zhang, K., & Aslan, A (2021). AI technologies for education: Recent research & future directions. Computers and Education: Artificial Intelligence, 2, 1-11. https://www.sciencedirect.com/science/article/pii/S2666920X21000199?ref=pdf_download&fr=RR-2&rr=8dd94b389f77ade3 [↩]
12. Salas-Pilco, S. & Yang, Y (2022). Artificial intelligence applications in Latin American higher education: a systematic review. International Journal of Educational Technology in Higher Education,19 (21), (2022). https://doi.org/10.1186/s41239-022-00326-w [↩] [↩] [↩] [↩] [↩] [↩] [↩] [↩]
13. Drèze, J., & Sen, A. (2013). An Uncertain Glory: India and its contradictions. Princeton University Press [↩]
14. PROBE Team. (1999). Public report on basic education in India. Oxford University Press [↩]
15. UNESCO. (2020). Sustainability education: A roadmap. https://doi.org/10.54675/YFRE1448 [↩]
16. Oxfam. (2016). Even it up: Time to end extreme inequality. Oxfam International [↩]
17. UNESCO. (2003). Education in a multilingual world. https://unesdoc.unesco.org/ark:/48223/pf0000129728 [↩]
18. UNESCO (2023). How can Artificial Intelligence enhance education. https://www.unesco.org/en/articles/how-can-artificial-intelligence-enhance-education [↩] [↩]
19. Ayeni, O., Al Hamad, N.M., Chisom, O.N., Osawaru, B. (2024). AI in education: A review of personalized learning and educational technology. GSC Advanced Research and Reviews, 18(02), 261–27. https://doi.org/10.30574/gscarr.2024.18.2.0062 [↩] [↩] [↩] [↩] [↩]
20. Hakimi, M. & Shahidzay, A. K. (2024) Transforming Education with Artificial Intelligence: Potential and Obstacles in Developing Countries. Preprints , 2024072542. https://www.researchgate.net/publication/383129232_Transforming_Education_with_Artificial_Intelligence_Potential_and_Obstacles_in_Developing_Countries [↩] [↩]
21. VanLehn, K. (2011). The relative effectiveness of human tutoring, intelligent tutoring systems, and other tutoring systems. Educational Psychologist, 46(4), 197-221. https://www.researchgate.net/publication/233237328_The_Relative_Effectiveness_of_Human_Tutoring_Intelligent_Tutoring_Systems_and_Other_Tutoring_Systems [↩] [↩]
22. Chisom, O., Unachukwu, C., & Osawaru, B. (2023). Review of AI in Education: Transforming Learning Environments in Africa. International Journal of Applied Research in Social Sciences, 5 (10), 637-654. https://fepbl.com/index.php/ijarss/article/view/725 [↩]
23. Bhutoria, A. (2022). Personalized education and Artificial Intelligence in the United States, China and India: A systematic review using a Human in the Loop model. Computers and Education: Artificial Intelligence, 3, 1-18. https://www.sciencedirect.com/science/article/pii/S2666920X22000236 [↩] [↩] [↩] [↩]
24. Moyo, Sichelesile, Nyoni, Phillip, Mutunhu Ndlovu, Belinda, Dzinomwa Mary, Dube Sibusisiwe and Dube, Sinokubekezela. Assessment of Students’ Experiences and Viewpoints in Using Chatbots for Education Practice: A University Case of a Developing Country (https://www.researchgate.net/publication/384240732_Assessment_of_Students’_Experiences_and_Viewpoints_in_Using_Chatbots_for_Education_Practice_A_University_Case_of_a_Developing_Country) [↩]
25. Moyo, Sichelesile, Nyoni, Phillip, Mutunhu Ndlovu, Belinda, Dzinomwa Mary, Dube Sibusisiwe and Dube, Sinokubekezela. Assessment of Students’ Experiences and Viewpoints in Using Chatbots for Education Practice: A University Case of a Developing Country, (https://www.researchgate.net/publication/384240732_Assessment_of_Students’_Experiences_and_Viewpoints_in_Using_Chatbots_for_Education_Practice_A_University_Case_of_a_Developing_Country) [↩]
26. Avelar, M. & Terway, A (2020). Zaya Learning Lab: A Case of Impact Investment. NORRAG, Switzerland. https://resources.norrageducation.org/resource/631/zaya-learning-lab-a-case-of-impact-investment [↩]
27. Educational Initiatives (2024). https://ei.study/ [↩] [↩]
28. Educational Initiatives (2024). https://ei.study/ [↩]
29. Jet Education Services (2024). https://www.jet.org.za/ [↩] [↩] [↩] [↩] [↩]
30. Tahiru, F. AI in Education: A Systematic Literature Review. Journal of Cases on Information Technology, 23 (1), 1-20.https://www.igi-global.com/article/ai-in-education/266434 [↩] [↩]
31. Trucco, D., & Palma, J. (2021). Digital technologies in Latin America and the Caribbean: What do we know and what do we need to know? Inter-American Development Bank [↩]
32. Thomson, R. & Anderson, J. (2023). Digital Education Revolution: Navigating the Challenges of AI Integration in Developing Nations, 1-9. https://www.researchgate.net/publication/376140219_Digital_Education_Revolution_Navigating_the_Challenges_of_AI_Integration_in_Developing_Nations [↩] [↩]
33. Zhang, K., & Aslan, A (2021). AI technologies for education: Recent research & future directions. Computers and Education: Artificial Intelligence, 2, 1-11. https://www.sciencedirect.com/science/article/pii/S2666920X21000199?ref=pdf_download&fr=RR-2&rr=8dd94b389f77ade3 [↩]
34. Aderibigbe, A., Ohenhen, P., Nwaobia, N., Gidiagba, J., Ani, E. (2023). Artificial Intelligence in Developing Countries: Bridging the Gap between Potential and Implementation. Computer Science & IT Research Journal, 4(3), 185-199. https://www.researchgate.net/publication/376196753_ARTIFICIAL_INTELLIGENCE_IN_DEVELOPING_COUNTRIES_BRIDGING_THE_GAP_BETWEEN_POTENTIAL_AND_IMPLEMENTATION [↩]