EdTech Archives EdTech Archives The Journal of Applied Instructional Design, 15(2)

The Integrated Ecosystem Approach to AI-Enabled Knowledge Infrastructure for Low-Connectivity Contexts

Hyojung Kim, Shouhong Zhang, & Sandip Bordolio

Abstract

AI deployments often fail in underserved regions because tools are shipped without the knowledge infrastructure required to operate under severe connectivity constraints. We present an integrated ecosystem approach to AI-enabled knowledge infrastructure, operationalized via TrueLeap solutions across schools in the Democratic Republic of Congo (DRC). The approach couples edge AI for localization and personalization, multilingual content pipelines, teacher co-pilot tools and analytics, and local capacity building and addresses three systemic gaps: (1) sample representation; (2) inclusivity & transparency (privacy-first and locally tuned models instead of generic one-size-fits-all plugins); and (3) workforce diversity (teacher PD and youth pipelines from digital literacy to advanced computing). We contribute: (i) a practical architecture for knowledge infrastructure in low-connectivity contexts; (ii) governance patterns that operationalize responsible AI; and (iii) evidence that human facilitation + edge AI advances equitable outcomes under infrastructural constraints.

Introduction

With the advent of new technologies and burgeoning industries that require a more competent workforce, teaching and learning practices have been, and are still changing in a form that is more innovative, borderless, and open. While a new form of learning has been identified as integral to preparing students for the paradigm shift, few efforts have been made to understand how these changes are occurring in emerging markets in the Global South. Moreover, many innovative attempts to transform education have failed to recognize the importance of community-anchored and contextually adaptive approaches in implementing new educational technology (Koul & Nayar, 2021).

A larger number of previous studies investigated how a particular educational technology has been or should be implemented in emerging markets (Animashaun et al., 2024; Lambrechts et al., 2020; Wong & Sixl-Daniell, 2017). However, much of the existing work has been conducted within a fragmented approach. Fragmented approaches that prioritize the perspectives of a single stakeholder or a limited set of stakeholders offer an incomplete understanding of the complex dynamics required for effective multi-actor collaboration in implementation processes. While the holistic ecosystem approach seeks to address these gaps, it has not yet been widely tested in diverse contexts (Koul & Nayar, 2021; Norris et al., 2013; Shaikh, 2023).

Thus, to address this gap in research and practice, a holistic ecosystem approach to the use of educational technology in emerging markets is needed. Such an approach is particularly valuable for deepening understanding of how educational technology is implemented in practice from administrative, academic, and student-centered perspectives. By integrating the diverse perspectives of multiple stakeholders, the ecosystem approach also holds potential to be more sustainable, scalable, and cost-effective.

Literature Review

Existing Fragmented Approaches

Several initiatives have sought to address educational changes and challenges in the Global South through technological interventions. However, many of these efforts have adopted fragmented approaches that focus on solving specific problems rather than integrating new technology into broader social and contextual practices. For example, a technological intervention was deployed to address a lack of knowledge and skills in basic literacy and arithmetic among the Mexican student population (Vila-Rosado et al., 2017). To solve this problem, innovative educational technology, a computer that includes an OS system, was introduced, and the research team proceeded to implement it in local schools. While the main goal of the project was to provide students with improved access to internet resources even without a direct connection, the approach placed less emphasis on integrating new teaching methods and pedagogical practices that aligned with the technology. In addition, limited attention was given to the software component, the systematic tracking of students’ academic progress, and the development of adequate tools to measure learning outcomes following the adoption of computers. Similarly, Iyamu and Ogiegbaen (2005) examined the use of various educational technologies in social studies classes in Nigeria, highlighting concerns about students’ creativity, critical thinking, and problem-solving skills. Their study aimed to explore how technology could be enhanced to support these essential skills for the future. The research primarily gathered insights from educators by surveying teachers about the frequency of technology use and factors influencing their adoption. While this focus provided valuable information on teacher perspectives, it did not include input from students, school administrators, or technology providers, which may limit a comprehensive understanding of the factors affecting technology integration and usage in the classroom.

Building on these observations, many existing solutions tend to address isolated components of the problem rather than offering a comprehensive strategy. For instance, Abrami et al. (2016) conducted an empirical study on the implementation of multimedia literacy software in early elementary schools in Kenya. The study highlighted effective teacher training and reported that students using the software demonstrated improved literacy skills; however, it placed less emphasis on developing a holistic administrative and technological strategy to scale the intervention to a broader learner population.

Challenges in Emerging Markets

As illustrated by the previous examples, the transformative potential of technology in education remains largely unrealized. Large-scale adoption and seamless integration of new educational technologies continue to face significant challenges in the Global South. Efforts to introduce and adapt technology within existing teaching and learning environments are often fragmented and difficult to sustain (Norris et al., 2013). This ongoing underutilization of educational technology in emerging markets is increasingly recognized as a critical issue, especially given technology’s important role in equipping students with fundamental literacy, numeracy, and digital skills

Researchers have been investigating both the systemic and technical factors that pose challenges to the adoption of educational technology in emerging markets. Naresh and Reddy (2015) suggest that a lack of awareness, a lack of a systematic approach to technology adoption, and difficulty transforming the educational system pose challenges. Similarly, Gulati (2008) highlights the difficulty of distributing new technology to larger populations in emerging economies because certain groups cannot afford the technology or the education required. Another important factor that contributes to challenges is the socio-cultural landscape of emerging markets. According to Nawaz and Qureshi (2010), adaptiveness, receptiveness, and perception towards new technology depend on the population’s prior experience, age, gender, and culture. Thus, a technology adoption model widely accepted in advanced economies may cause difficulties and disconnect the technology from local culture.

While a significant amount of research has focused on analyzing and generalizing the factors influencing technology adoption, some case studies have targeted specific markets and provided rich, realistic descriptions of the challenges involved. For example, Sife et al. (2007) conducted a case study on ICT implementation in Tanzanian universities. The results presented 6 factors that contribute to the challenges in ICT implementation, which are 1) lack of systematic approach, 2) lack of awareness and attitude towards ICT, 3) lack of administrative support, 4) lack of technological support, 5) lack of ownership, and 6) inadequate funds. Along similar lines, Oroma et al. (2012) identified the risk factors for implementing e-learning technology in Uganda. According to the analysis, the four critical challenges against adopting ICT in developing countries are 1) inequality in access, 2) the lack of and the high cost of equipment, 3) the lack of skills and training on the technology, and 4) the lack of pedagogy and trained teachers. Taurus et al. (2015) similarly analyzed the factors that impeded the introduction of e-learning solutions in public universities in Kenya. It was found that 1) inadequate ICT and e-learning infrastructure, 2) lack of funds, 3) lack of affordable and sustainable internet bandwidth, 4) lack of operational e-learning policies, 5) lack of teacher training on the new technology, and 6) lack of interest and commitment among the teachers acted as significant barriers for implementing new educational technologies. Synthesizing the challenges identified across the reviewed studies reveals that inadequate teacher training, institutional constraints such as limited funding, and insufficient access to advanced technological infrastructure constitute significant barriers to technology adoption. Additionally, the absence of a context-sensitive, integrated approach that accounts for the diverse conditions characteristic of emerging markets further impedes successful implementation.

Responsible AI from Ethics to Action

We frame three ethical gaps and practical responses that anchor the ecosystem approach. First is the sample representation and data poverty issue. Mainstream LLM corpora are heavily English-centric. For example, the C4 corpus used for T5 is explicitly English only (Dodge et al., 2021), and in a similar manner, an analysis of the Common Crawl CC-MAIN-2023-50 snapshot was found to be approximately 44% in English, with only 14 of 160 languages exceeding 1% share (Perełkiewicz & Poświata, 2024). This over-representation of the English language in the existing AI tools propagated into popular models. The results are demonstrated in LLaMA’s pretraining mix, including 67% English CommonCrawl and ~15% English C4 (Touvron et al., 2023). Furthermore, the connectivity issue compounds the data privacy issue. To elaborate, in 2024, only 27% of people in low-income countries were reported to have access to the internet. Similarly, 5G coverage in low-income countries was less than 4%, and mobile internet costs in Africa were estimated at 14 times the average European level (ITU, 2024). According to GSAM (2024), despite widespread mobile broadband coverage, 39% of the world, which accounts for approximately 3.1 billion people, live under coverage but do not use mobile internet. This phenomenon, referred to as the ‘usage gap’, is now nine times higher than the size of the coverage gap (GSMA, 2024). Participatory, multilingual data practices and community-driven pipelines are widely recommended to redress these representation gaps (Nekoto et al., 2020; UNESCO, 2023). The proposed ecosystem approach invests in local digital infrastructure (e.g., technology labs and edge capture), multilingual data pipelines, and participatory data practices, so models reflect local languages, dialects, and contexts.

The second ethical gap is the inclusivity, localization, and privacy issues. Generic, one-size-fits-all plugins often encode hidden assumptions that misfit local pedagogies and sociocultural norms, creating inequities and safety risks when exported to new contexts (Sambasivan et al., 2021). To mitigate this, we favor locally tuned models and documentation practices that make dataset/model assumptions explicit and traceable (Bender & Friedman, 2018; Mitchell et al., 2019). We implement privacy-by-design—data minimization, consent, de-identification, and regional data sovereignty—aligned with established risk-management guidance (Cavoukian, 2011; NIST, 2023). Because many deployments operate under severe bandwidth constraints, we support real-time interactions via edge/offline execution with store-and-forward synchronization (Hamdan et al., 2020). Finally, we operationalize safety at the edge—running sensitive inference on-device/locally hosted to minimize data exposure—while enforcing offline-first controls (signed model updates, local decision logs, delayed-sync audit trails). This implements privacy-by-design and risk-managed deployment in bandwidth-constrained settings (Hamdan et al., 2020; Cavoukian, 2011; NIST, 2023).

The third gap is establishing workforce diversity at scale. A concentrated global AI workforce limits perspectives and can perpetuate bias; our ecosystem, therefore, pairs teacher professional development with youth talent pipelines (digital literacy → advanced computing) to cultivate a diverse AI workforce across regions. This aligns with global guidance calling for inclusion and diverse participation in AI design and deployment (UNESCO, 2023). To scale sustainably, we emphasize Train-the-Trainer (ToT) cascades that localize expertise and reduce reliance on external specialists—an approach recommended for AI in education and supported by evidence of cascade effectiveness in low-resource settings (UNESCO IITE, 2022; Zafar et al., 2016). We also recognize the importance of leadership and management training for school and system leaders, so that institutions can plan, resource, and monitor AI-enabled learning at scale (UNESCO IESALC, n.d.; OECD, 2023). Finally, the youth pipeline is anchored in recognized digital/AI literacy frameworks to ensure progression from basic skills to advanced computing (UNICEF, 2020; OECD, 2023). These pillars translate ethical intent into programmatic requirements for technology, governance, and human capacity inside the knowledge infrastructure.

Ecosystem Approach Conceptualization

Theoretical foundation

A holistic ecosystem approach is a perspective that acknowledges and addresses the fact that educational innovation extends beyond the mere adoption of new technology; it requires adapting these technologies to diverse classroom environments shaped by varying cultural, economic, and geographic contexts. Moreover, this framework is essential for addressing the distinct needs of multiple stakeholders, including teachers, students, institutions, and policymakers. According to Koul and Nayar (2021), the Holistic Learning Educational Ecosystem identifies the roles of relevant stakeholders in both education and industry while keeping students and learning at the center. Similarly, Norris et al. (2013) define the educational ecosystem as one in which members benefit from each other’s participation through symbiotic relationships. From this perspective, any type of educational technology becomes a digital environment populated by living species that interact, exhibit independent behavior, and evolve (Uden et al., 2007, p. 114). Each component of the ecosystem—analogous to a living species—warrants equal consideration and evaluation to enhance overall efficiency, effectiveness, and innovation. By ensuring balanced attention to all stakeholders, this approach fosters a sustainable learning environment that maximizes positive outcomes for both learners and institutions.

In the ecosystem, a co-value is created and shared among the related stakeholders, fostering a symbiotic relationship within the broader context so that the shared interest in enhancing learning can continue to resonate, ultimately channeling every contributor towards a unified goal (Koul & Nayar, 2021). Having a unified goal and each stakeholder contributing to the co-value is at the core of the holistic ecosystem approach. Within this approach to viewing the classroom setting as an ecosystem, multiple contributors, working towards their own goals, directly and indirectly contribute to the institution’s broader vision. Adopting this framework, Koul and Nayar (2021) delineates 10 drivers to employ this approach into realistic classroom settings, which includes 1) unique student experiences, 2) experiential learning, 3) smart classrooms, 4) global connect, 5) cost-effective teaching pedagogies, 6) knowledge/research bank creation, 7) peer to peer learning, 8) faculty as a facilitator, coach, and a mentor, 9) institutes as a resource facilitator, and 10) policy-makers as guiding architects.

This holistic approach has profound implications for various stakeholders and, by involving all parties, ensures a more integrated and effective adoption of new educational technology. The first step in implementing this holistic ecosystem approach is to shift the focus from technology adoption to actual adaptation by examining the problems that students, teachers, and staff face and addressing administrative and technical frustrations (Norris et al., 2013).

Stakeholder Components

Contrary to the fragmented approach, the holistic ecosystem approach in educational technology implementation actively aims to incorporate the varying aspects of different stakeholders. Thus, this study identifies 4 representative stakeholders most frequently included in the ecosystem approach and delineates their respective roles. The 4 critical stakeholders are 1) learners, 2) educators, 3) policymakers, and 4) institutions.

For learners, the primary benefit of the ecosystem approach lies in the creation of personalized learning experiences. With the aid of technology, education can be tailored to meet individual learning needs and allow students to progress at their own pace. This personalization enables learners to engage with content that suits their learning style and ultimately take ownership of their educational journey. This level of customization helps address diverse learning abilities and encourages greater engagement and knowledge retention.

For educators, the role of the teacher should be expanded beyond traditional lecturing to include facilitation and mentoring within the ecosystem approach. Technology empowers educators to guide students through interactive and dynamic learning experiences, fostering critical thinking, collaboration, and problem-solving. Educators are also given the opportunity to use data and analytics to better understand student performance and adjust instruction accordingly, leading to more effective teaching strategies. Educators may choose to conduct face-to-face sessions with all students in the classroom or facilitate independent learning by assigning real-world problem-solving tasks supported by technological tools (Oroma et al., 2012). Consequently, educators are pivotal stakeholders in the adoption and effective use of educational technology.

Policymakers also play a crucial role in shaping the direction of educational technology by adapting curriculum development to meet the evolving needs of the workforce. This requires the ability to design flexible, forward-thinking curricula that incorporate new technologies and teaching methodologies. Policymakers must ensure that education systems can respond rapidly to technological and industry changes, equipping students with the skills needed to thrive in a digital economy.

Institutions

Finally, institutions must provide comprehensive infrastructure support to facilitate the successful integration of educational technology. This includes not only the provision of hardware and software but also reliable internet access, ongoing technical support, and training for both educators and learners. Institutions play a key role in creating the necessary environment for educational technology to thrive, ensuring that both physical and digital infrastructure are in place to support the evolving educational landscape.

Technology Solution Components

The holistic ecosystem approach presented in this paper not only incorporates the diverse perspectives of stakeholders but also identifies the critical components of an effective software solution. Thus, according to the holistic ecosystem model, a new type of educational technology model is introduced to tailor to the hyper-local, multi-centric nature of technology integration. The new ecosystem approach for educational technology is proposed to satisfy the following five elements: 1) hardware infrastructure and connectivity, 2) AI-enabled software and learning community, 3) data-driven collective intelligence, 4) digital content, and 5) training & support.

AI-enabled software platform and e-learning community

An AI-enabled e-learning platform, including a Learning Management System (LMS) and a Content Management System (CMS), revolutionizes the traditional learning environment by personalizing education and adapting to each learner's unique needs. These platforms harness artificial intelligence to analyze student data, providing insights into individual learning patterns and performance. AI-driven tools can suggest tailored learning paths, recommend resources, and offer real-time feedback to learners. Additionally, AI can help identify learning gaps, enabling instructors to provide targeted interventions. The software should not only focus on individual learning, but also foster a learning community where students, educators, and even parents can interact and collaborate. This connected ecosystem creates a sense of belonging and encourages peer-to-peer learning, discussion forums, and group activities, allowing learners to engage with their peers in meaningful ways despite geographical boundaries. Additionally, the community aspect supports social-emotional learning by encouraging students to communicate, collaborate, and develop interpersonal skills in a digital environment.

Thus, the e-learning platform should facilitate the creation of a learning community where educators, learners, and parents can interact in real time. Virtual classrooms, discussion boards, and collaborative project tools are key features that enable learners to share insights, ask questions, and work together. Communities can also be built around specific subjects, interests, or goals, promoting a sense of support, engagement, and motivation.

Hardware Infrastructure and Connectivity

A robust and reliable hardware infrastructure is the backbone of any successful educational technology implementation. This involves ensuring that schools, institutions, and learners have access to the necessary devices, such as computers, tablets, or interactive whiteboards. Equally important is reliable and high-speed internet connectivity, which enables seamless access to digital content, online learning resources, and e-learning platforms. In many emerging markets, connectivity can be a significant challenge; thus, addressing these gaps is crucial for creating an inclusive and accessible learning environment. Ensuring that all stakeholders—learners, educators, and institutions—have access to the appropriate hardware and the internet is a fundamental step in promoting equitable educational opportunities.

Data-driven Intelligence

Educational technology that incorporates data-driven intelligence can provide invaluable insights into both the learning process and the effectiveness of teaching strategies. By collecting data on student behavior, engagement, performance, and progress, the system can generate analytics that help inform decision-making at various levels. Teachers can use these insights to adjust their instructional approaches, while administrators and policymakers can identify trends, challenges, and opportunities for improvement. Data-driven systems can also help predict student outcomes, allowing for early intervention for at-risk learners. In this way, data serves as a powerful tool to ensure that the educational experience is optimized, personalized, and responsive to the needs of all stakeholders.

Locally Relevant Digital Content

The quality and relevance of digital content are critical to the success of any e-learning platform. This content includes videos, quizzes, interactive lessons, simulations, and other multimedia elements that cater to different learning styles. The digital content should be engaging, up to date, and aligned with curriculum standards to ensure learners gain the necessary knowledge and skills. Content must also be adaptable, offering various levels of difficulty to cater to diverse learner needs. Moreover, it should be culturally sensitive and contextually relevant, especially in emerging markets where learners may have different cultural backgrounds, learning experiences, and access to resources.

Training and Support

The successful adoption of educational technology requires comprehensive training and support for all stakeholders involved. Educators must be equipped with the skills and knowledge to use technology effectively and integrate it seamlessly into their teaching practices. Ongoing professional development and training programs can help teachers stay up to date with the latest tools, pedagogies, and best practices. Similarly, learners must be provided with training on how to navigate the platform and make the most of its features. This may include tutorials on accessing learning materials, communicating with peers, and using assessment tools. Additionally, technical support should be readily available for both learners and educators to resolve any issues they may encounter with the platform or its features. Ensuring that both students and teachers feel confident in using educational technologies is crucial to maximizing the platform’s potential and ensuring its long-term success.

Methodology

As part of the research, the team examined the implementation of an integrated educational technology ecosystem, deploying TrueLeap’s knowledge-infrastructure solution with Akili Digital, the Catholic School Association, and the Parents’ Association across seven schools serving an estimated 10,000 students at the time of research in the Democratic Republic of Congo (DRC). When implementing a new educational technology in the Global South within the ecosystem framework, the initiative should be guided by three important goals: affordability, sustainability, and scalability.

Affordability

First of all, the researchers implemented measures to maintain and protect the affordability of implementing a new tool in the Global South. To carry out this principle, the AI-enabled software platform and hardware infrastructure are developed to be cost-effective and purpose-fit. For affordability, it is also important to reduce the costs of integrating disparate systems from multiple vendors and to reduce replacement frequency. These have been accomplished by providing an integrated solution that encompasses hardware, software, content, and training; by designing the system to perform effectively on older hardware specifications; and, more importantly, by training local talent to provide timely support and maintenance.

Scalability

Second, in terms of scalability, the platform employs a modular architecture allowing schools and educational systems to begin with basic functionality and expand capabilities as needed. This "start small, grow big" approach can be deployed in diverse settings, from urban to deeply rural environments, through adaptable implementation approaches, from fully online to primarily offline systems with periodic synchronization.

Another critical circumstance for protecting scalability for any educational measure is standardizing implementation frameworks for further adoption to new contexts. For this context of scalability, the knowledge infrastructure solution has been developed into replicable implementation models that can be efficiently adapted to new contexts.

Sustainability

Last but not least, in terms of sustainability, the researchers have designed the e-learning platform to promote localized content and cultural integration. The platform prioritizes local relevance by supporting multiple languages, including regional dialects. In implementation cases like DRC, the interface incorporates not only French and English but also local languages such as Lingala, Swahili, Kikongo, and Tshiluba. This linguistic inclusivity ensures the solution remains relevant and embedded in local educational ecosystems.

To promote sustainability, the knowledge infrastructure solution also recognizes the need to provide offline functionality and low resource requirements. For example, the Learning Management System (LMS) is engineered to function effectively on modest hardware specifications and limited bandwidth, making it viable even in resource-constrained environments without requiring constant hardware upgrades. Another aspect of ensuring the sustainability of educational technology interventions in the Global South is establishing local training and support ecosystems. To address this issue, rather than relying on foreign technical support, the comprehensive system established local training programs and support networks.

Expected Outcomes

The new integrated ecosystem approach to knowledge infrastructure is expected to impact across multiple dimensions among various stakeholders. Based on in-depth on-the-ground interviews, the initial implementation is already enhancing learning environments, improving technological access, fostering positive stakeholder engagement, and establishing a sustainable implementation model that transcends traditional technological interventions.

Enhanced learning environments

First, the integrated ecosystem approach to introducing new educational technology in the Global South is expected to significantly enhance learning environments, thereby improving the student experience in terms of engagement, motivation, and mental well-being. This gain is demonstrated in the testimonials from one of the school teachers, reporting: “Students did have a chance to be trained by using the new technology. I am excited to see them in front of the machines. Our children have six hours of class per week, so for some time they did not have the means to use the computer, but now they can get the training.”

Affordable technological access

Second, the integrated ecosystem approach is providing a low-connectivity community with affordable, readily accessible access to technology. This is demonstrated in the teacher’s testimonial in DRC reporting, “Other systems required us to adapt our teaching to their design. This system adapts to how we teach and what our students need.” The Minister of Primary, Secondary, and Technical Education of DRC also reinforced this finding by stating, “My presence is a testament to my support and encouragement. Together, we aim to elevate the quality of teaching and learning for our children. Thanks to the technology, achieving optimal educational conditions is now possible in our beautiful country.”

Positive stakeholder engagement

Third, the integrated ecosystem approach is promoting positive stakeholder engagement. This is demonstrated by the testimonial of Monsieur Abbe N’Tungu Bisibu Noel, the Head of the Catholic School Association of DRC, reporting, “We are working together for schools to improve learning tools for the children, and improve the quality of education in the Democratic Republic of Congo.” The stakeholder benefits of this holistic approach extend to the parent community's feedback. For example, one parent in DRC noted that "We can now see what our children are learning. Before, we had no visibility into the classroom." This reveals that the holistic ecosystem approach fosters greater trust in the educational system by distributing benefits to the broader student-parent community.

Establishment of a Sustainable Model

Lastly, the integrated ecosystem approach is making a significant impact in generating and distributing a sustainable, scalable model. One critical element in ensuring the long-term scalability and sustainability of educational technology is establishing well-trained staff and personnel by providing local talent with high-fidelity professional development and training opportunities. Also, school administrators are experiencing a significant reduction in time spent on manual record-keeping and reporting, allowing the reallocation of staff resources to student support functions.

Discussion and Conclusions

The findings of this study highlight the critical role of an integrated ecosystem approach in addressing the challenges emerging markets face in adopting new educational technology. The case study in DRC demonstrates how the inclusion of multiple stakeholders—educators, students, policymakers, and institutions—can create a sustainable and scalable model for technology adoption. The research shows that the effective use of educational technology is not merely about introducing new tools but about adapting those tools to the specific needs and contexts of diverse classrooms.

The results also underscore the necessity of infrastructure renovation, tailored training, and localized content to ensure that technological tools meet the unique needs of both students and teachers. The significant improvement in student engagement, motivation, and performance observed in the case study further supports the value of a comprehensive approach that connects various stakeholders and integrates their diverse perspectives. Furthermore, the study suggests that incorporating community-based approaches, in which local stakeholders take ownership of the technology, is vital to ensuring that innovations are not only adopted but also sustained over time.

In conclusion, this study underscores the importance of an integrated ecosystem approach to educational technology adoption in emerging markets. The key takeaway from this research is that technology adoption must be seen not as a one-time solution, but as an ongoing process of adaptation, training, and collaboration among various stakeholders. Finally, the findings contribute to the growing body of literature on educational technology adoption by emphasizing the value of a holistic, ecosystem-based model that can be adapted and scaled across regions to address the unique challenges emerging markets face in the AI age.

References

  1. Abrami, P. C., Wade, C. A., Lysenko, L., Marsh, J., & Gioko, A. (2016). Using educational technology to develop early literacy skills in Sub-Saharan Africa. Education and Information Technologies, 21, 945–964.
  2. Animashaun, E. S., Familoni, B. T., & Onyebuchi, N. C. (2024). Implementing educational technology solutions for sustainable development in emerging markets. International Journal of Applied Research in Social Sciences, 6(6), 1158–1168.
  3. Bender, E. M., & Friedman, B. (2018). Data statements for natural language processing: Toward mitigating system bias and enabling better science. Transactions of the Association for Computational Linguistics, 6, 587–604. https://doi.org/10.1162/tacl_a_00041
  4. Cavoukian, A. (2011). Privacy by design: The 7 foundational principles—Implementation and mapping of fair information practices. Information and Privacy Commissioner of Ontario.
  5. Dodge, J., Ilharco, G., Schwartz, R., Farhadi, A., Hajishirzi, H., Smith, N. A., & Gardner, M. (2021). Documenting large webtext corpora: A case study on the Colossal Clean Crawled Corpus. Proceedings of EMNLP 2021, 1286–1305. https://doi.org/10.18653/v1/2021.emnlp-main.98
  6. Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé III, H., & Crawford, K. (2021). Datasheets for datasets. Communications of the ACM, 64(12), 86–92. https://doi.org/10.1145/3458723
  7. GSMA. (2024). The State of Mobile Internet Connectivity Report 2024. GSMA.
  8. Gulati, S. (2008). Technology-enhanced learning in developing nations: A review. International Review of Research in Open and Distributed Learning, 9(1), 1–16.
  9. Hamdan, S., Ayyash, M., & Almajali, S. (2020). Edge-computing architectures for Internet of Things applications: A survey. Sensors, 20(22), 6441.  https://doi.org/10.3390/s20226441
  10. International Telecommunication Union (ITU). (2024). Measuring digital development: Facts and figures 2024. ITU.
  11. Iyamu, E. O., & Ogiegbaen, S. (2005). Assessment of the use of educational technology by social studies teachers in secondary schools in western Nigeria. Australian Journal of Educational Technology (AJET), 8(1).
  12. Koul, S., & Nayar, B. (2021). The holistic learning educational ecosystem: A classroom 4.0 perspective. Higher Education Quarterly, 75(1), 98–112.
  13. Lambrechts, W., Sinha, S., & Marwala, T. (2020). Decentralizing emerging markets to prepare for Industry 4.0: Modernizing policies and the role of higher education. In The disruptive Fourth Industrial Revolution: Technology, society and beyond (pp. 111–153).
  14. Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., Spitzer, E., Raji, I. D., & Gebru, T. (2019). Model cards for model reporting. In Proceedings of the 2019 Conference on Fairness, Accountability, and Transparency (pp. 220–229). ACM. https://doi.org/10.1145/3287560.3287596
  15. Naresh, B., & Reddy, B. S. (2015). Challenges and opportunity of E-learning in developed and developing countries—A review. International Journal of Emerging Research in Management & Technology, 4(6), 259–262.
  16. Nawaz, A., & Qureshi, Q. A. (2010). E-teaching/E-pedagogy: Threats & opportunities for teachers in HEIs. Global Journal of Management & Business Research, 10(9), 23–31.
  17. Nekoto, W., Marivate, V., Matsila, T., Fasubaa, T., Kolawole, T., Fagbohungbe, T., … the Masakhane Community. (2020). Participatory research for low-resourced machine translation: Masakhane. Findings of EMNLP 2020, 438–455. https://doi.org/10.18653/v1/2020.findings-emnlp.372
  18. Norris, L., Eyt-Dessus, A., & Holtham, C. (2013). The learning ecosystem: A practical, holistic approach to old problems in a new world. In ASCILITE 2013 Conference Proceedings (pp. 633–641).
  19. OECD. (2023). OECD Digital Education Outlook 2023: Towards an Effective Digital Education Ecosystem. OECD Publishing. https://doi.org/10.1787/c74f03de-en
  20. Oroma, J., Wanga, H., & Fredrick, N. (2012). Challenges of e-learning in developing countries: The Ugandan experience. In INTED2012 Proceedings (pp. 3535–3543). IATED.
  21. Perełkiewicz, M., & Poświata, R. (2024). A review of the challenges with massive web-mined corpora used in large language models pre-training. arXiv preprint arXiv:2407.07630.
  22. Raji, I. D., Smart, A., White, R. N., Mitchell, M., Gebru, T., Hutchinson, B., Smith-Loud, J., Theron, D., & Barnes, P. (2020). Closing the AI accountability gap: Defining an end-to-end framework for internal algorithmic auditing. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency. https://doi.org/10.1145/3351095.3372873
  23. Saleiro, P., Kuester, B., Hinkson, L., London, J., Stevens, A., Anisfeld, A., Rodolfa, K. T., & Ghani, R. (2018). Aequitas: A bias and fairness audit toolkit. arXiv preprint arXiv:1811.05577.
  24. Shaikh, Z. (2023). Integration of digital pedagogy and development of holistic learning ecosystem for students: The teachers’ perspective. International Journal of Educational Reform. https://doi.org/10.1177/10567879231193729
  25. Sife, A., Lwoga, E., & Sanga, C. (2007). New technologies for teaching and learning: Challenges for higher learning institutions in developing countries. International Journal of Education and Development Using ICT, 3(2), 57–67.
  26. Tarus, J. K., Gichoya, D., & Muumbo, A. (2015). Challenges of implementing e-learning in Kenya: A case of Kenyan public universities. The International Review of Research in Open and Distributed Learning, 16(1). https://doi.org/10.19173/irrodl.v16i1.1816
  27. Touvron, H., Lavril, T., Izacard, G., Martinet, X., Lachaux, M.-A., et al. (2023). LLaMA: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
  28. Uden, L., Wangsa, I. T., & Damiani, E. (2007). The future of e-learning: E-learning ecosystem. 2007 Inaugural IEEE-IES Digital Ecosystems and Technologies Conference, 113–117. https://doi.org/10.1109/DEST.2007.371955
  29. UNESCO. (2023). Guidance for generative AI in education and research. UNESCO.
  30. UNESCO IESALC. (n.d.). Leadership, management and training programme for higher education. https://www.iesalc.unesco.org/en/development/leadership-management-training
  31. UNESCO IITE. (2022). Analytical report on the use of advanced ICT/AI for digital transformation of education. UNESCO Institute for Information Technologies in Education. https://iite.unesco.org/publications/analytical-report-on-the-use-of-advanced-ict-ai-for-digital-transformation-of-education/
  32. UNESCO. (2023). ICT Competency Framework for Teachers (Version 3). https://www.unesco.org/en/digital-competencies-skills/ict-cft
  33. UNICEF. (2020). Digital literacy for children: Exploring definitions and frameworks. UNICEF Office of Research – Innocenti.
  34. UNICEF. (2021). Policy guidance on AI for children. UNICEF Office of Global Insight and Policy.
  35. Vila-Rosado, D.-E., Alamán-Garcia, A., & Aguire-Franco, I. (2017). Endless effort to improve education in mathematics and sciences in emerging markets: A case study. INTED Proceedings, 1, 9763–9771. https://doi.org/10.21125/inted.2017.2315
  36. Wenzek, G., Lachaux, M.-A., Conneau, A., Chaudhary, V., Guzmán, F., Joulin, A., & Grave, E. (2020). CCNet: Extracting high quality monolingual datasets from web crawl data. In Proceedings of LREC 2020 (pp. 4003–4012).
  37. Wolfenden, F., Addey, C., Cruz, F., & Westbrook, J. (2024). A framework for strengthening teacher professional development systems with ICT. Foundation for Information Technology Education and Development.
  38. Zafar, S., Sikander, S., Hamdani, S. U., Atif, N., Akhtar, P., Nazir, H., Maselko, J., & Rahman, A. (2016). The effectiveness of technology-assisted cascade training and supervision of community health workers in delivering the Thinking Healthy Program—Study protocol for a randomized controlled trial. Trials, 17(1), 1–11.  https://doi.org/10.1186/s13063-016-1287-9
  39. Wong, A. O., & Sixl-Daniell, K. (2017). The importance of e-learning as a teaching and learning approach in emerging markets. International Journal of Advanced Corporate Learning (iJAC), 10(1), 45. https://doi.org/10.3991/ijac.v10i1.6471
  40. National Institute of Standards and Technology. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0) (NIST AI 100-1). NIST.  https://doi.org/10.6028/NIST.AI.100-1
  41. Sambasivan, N., Arnesen, E., Hutchinson, B., Prabhakaran, V., & Birhane, A. (2021). Re-imagining algorithmic fairness in India and beyond. In FAccT ’21: Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 315–328). https://doi.org/10.1145/3442188.3445896

Acknowledgement

This study was made possible by support from TrueLeap Inc., which funded the research and provided access to de‑identified interview materials and operational data used in the analysis. We extend our sincere thanks to our partners in the Democratic Republic of Congo—Akili Digital, the Catholic School Association of the DRC, and the Parents’ Association—for their collaboration, field coordination, and sustained community engagement. We also thank Dr. Curt Bonk (Indiana University) for his support throughout this work. We are deeply grateful to the participating school leaders, teachers, learners, and families for generously sharing their time, perspectives, and experiences. Any opinions, findings, conclusions, or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of TrueLeap Inc. or our partners.