Digital Architectures in Education: Digital Architectures in Education: From Content Distribution Platforms to Adaptive Learning Spaces
The expansion of online and hybrid learning has significantly transformed educational practices in recent years. Massive Open Online Course (MOOC) platforms have been among the most visible manifestations of this transformation, providing large-scale access to standardized course materials and enabling mass participation. In many contexts, the use of MOOCs has been associated with improvements in academic performance and student satisfaction (Zhang et al., 2025). From this perspective, MOOCs have played an important role in democratizing access to education.
However, expanded access does not automatically lead to deep learning. Research indicates that merely completing a MOOC does not necessarily foster the development of self-regulation or self-efficacy (Wang & Zhu, 2019). Self-regulation involves the learner’s capacity to plan, monitor, and evaluate their own learning process (Zimmerman, 2002), and these competencies are not inherently developed through following a predefined instructional pathway. Furthermore, the literature highlights structural limitations such as excessive standardization and limited personalization (Ortega-Sánchez & Gómez-Trigueros, 2020).
In this context, the emergence of tools that go beyond content distribution and actively support learner engagement becomes particularly relevant. NotebookLM represents such an innovation. Unlike MOOCs, which provide a uniform curriculum for all participants, NotebookLM functions as a personalized learning environment built upon user-uploaded resources. The system can generate summaries, progressive explanations, and contextualized questions (Alisoy, 2025; Yeo et al., 2025), transforming reading into an active process of conceptual exploration.
From a theoretical standpoint, the distinction may be framed as follows: the MOOC represents a model of distributed instruction, whereas NotebookLM configures a model of personalized cognitive mediation. MOOCs optimize access; NotebookLM optimizes cognitive processing. While the former prioritizes scalability and standardization, the latter emphasizes personalization and AI-assisted dialogic interaction.
This distinction becomes evident in the way deep understanding is supported. The ability to request explanations at varying levels of complexity allows information to be adjusted to learners’ cognitive needs. From the perspective of Cognitive Load Theory (Sweller, 2011), progressive clarification can reduce cognitive overload and facilitate more effective knowledge organization. At the same time, interaction through structured questioning and response encourages active engagement, in contrast to the linear consumption of video-based content.
NotebookLM may also support self-regulation by generating formative questions and summaries, aligning with the principles of formative assessment, which emphasize feedback oriented toward learning progress rather than sanction (Black & Wiliam, 1998). However, the literature cautions against the risk of excessive delegation of analytical processes to the system (Albrecht-Crane, 2025). The educational value of such tools therefore depends on their critical and reflective use.
Another important dimension is adaptivity. Adaptive learning involves adjusting content according to the learner’s progress and needs (Peng et al., 2019). In the tradition of adaptive systems (Brusilovsky, 2001), personalization entails identifying knowledge gaps and providing differentiated support. In the case of NotebookLM, this may include adjusting question difficulty, suggesting additional resources, and revisiting insufficiently understood concepts. Nevertheless, effectiveness depends not only on technological capabilities but also on how these tools are embedded within a coherent pedagogical strategy (Rincon-Flores et al., 2024).
At the same time, potential limitations must be acknowledged. Language models operate through statistical prediction and may produce oversimplifications or distortions of meaning (Albrecht-Crane, 2025). There is also a risk of cognitive dependency, reduced analytical effort, and diminished critical autonomy. Issues related to data governance, algorithmic transparency, and educational equity remain central concerns. Inequalities in access to technology, already documented in the context of MOOCs (Ortega-Sánchez & Gómez-Trigueros, 2020), may be further amplified if not addressed systematically.
The MOOC, as a model of distributed instruction, and NotebookLM, as a model of personalized cognitive mediation, should not be regarded as competing solutions but as complementary layers within the same educational architecture. The future development of digital education does not lie in choosing one over the other, but in coherently integrating them: scalable infrastructure to ensure broad access, alongside personalization tools that support deep learning. This complementarity is fundamentally pedagogical and strategic, rather than merely technological.
References
Albrecht-Crane, C. (2025). Thinking smarter, not harder? Google NotebookLM’s misalignment problem in education. In Proceedings of the 43rd ACM International Conference on Design of Communication (SIGDOC ’25). Association for Computing Machinery. https://doi.org/10.1145/3711670.3764628
Alisoy, H. (2025). Can NotebookLM support English language learners? A theoretical perspective on AI tools in education. Porta Universorum, 1(6), 25–55. https://doi.org/10.69760/portuni.0106003
Black, P., & Wiliam, D. (1998). Assessment and classroom learning. Assessment in Education: Principles, Policy & Practice, 5(1), 7–74. https://doi.org/10.1080/0969595980050102
Brusilovsky, P. (2001). Adaptive hypermedia. User Modeling and User-Adapted Interaction, 11(1–2), 87–110. https://doi.org/10.1023/A:1011143116306
Ortega-Sánchez, D., & Gómez-Trigueros, I. M. (2020). MOOCs and NOOCs in the training of future geography and history teachers: A comparative cross-sectional study based on the TPACK model. IEEE Access, 8, 4035–4047. https://doi.org/10.1109/ACCESS.2019.2963314
Peng, H., Ma, S., & Spector, J. M. (2019). Personalized adaptive learning: An emerging pedagogical approach enabled by a smart learning environment. Smart Learning Environments, 6, Article 9. https://doi.org/10.1186/s40561-019-0089-y
Rincon-Flores, E. G., Castano, L., Guerrero Solis, S. L., Olmos Lopez, O., Rodríguez Hernández, C. F., Castillo Lara, L. A., & Aldape Valdés, L. P. (2024). Improving the learning-teaching process through adaptive learning strategy. Smart Learning Environments, 11, Article 27. https://doi.org/10.1186/s40561-024-00314-9
Sweller, J. (2011). Cognitive load theory. In J. P. Mestre & B. H. Ross (Eds.), Psychology of learning and motivation (Vol. 55, pp. 37–76). Academic Press. https://doi.org/10.1016/B978-0-12-387691-1.00002-8
Wang, K., & Zhu, C. (2019). MOOC-based flipped learning in higher education: Students’ participation, experience and learning performance. International Journal of Educational Technology in Higher Education, 16, Article 33. https://doi.org/10.1186/s41239-019-0163-0
Yeo, M. A., Moorhouse, B. L., & Wan, Y. (2025). From academic text to talk show: Deepening engagement and understanding with Google NotebookLM. English Teaching Journal, 28(4). https://doi.org/10.55593/ej.28112int
Zhang, Y., Liu, S., Li, C., & Hao, A. (2025). Impact of a MOOC-based learning model on performance, satisfaction, and perceptions in a histology course for Chinese medical students. Advances in Medical Education and Practice, 16, 1501–1510. https://www.tandfonline.com/doi/abs/10.2147/AMEP.S534676
Zimmerman, B. J. (2002). Becoming a self-regulated learner: An overview. Theory Into Practice, 41(2), 64–70. https://doi.org/10.1207/s15430421tip4102_2