Metaphysics and Metacognition in AI-Mediated Learning Environments: An Interpretivist Framework for Inclusive and International Education
Introduction
Artificial intelligence (AI) is now
deeply integrated into educational technologies, transforming pedagogical
practices, learner experiences, and institutional priorities. Prevailing
discussions about AI in education frequently focus on efficiency, personalisation,
and performance optimisation (Selwyn, 2019; Williamson et al., 2020). Although
these instrumental concerns are significant, they may obscure more fundamental
philosophical questions about the nature of learning, knowledge, and agency in
AI-mediated environments. In response, this paper proposes a theoretical
framework that foregrounds the metaphysical and metacognitive dimensions of
AI-supported learning.
This paper contends that AI-mediated
learning environments fundamentally alter the ontological and epistemological
foundations of education by redistributing cognition between human and
non-human actors. In this transformed context, metacognition is not only a
self-regulatory skill but also an ontological competence that enables learners
to interpret, navigate, and critically engage with AI systems. Grounded in an
interpretivist qualitative paradigm, this analysis situates these theoretical
concerns within inclusive classrooms and international school contexts, where
issues of diversity, epistemic authority, and learner agency are especially
salient.
Metaphysical
Reconfigurations of Learning in AI Contexts
Ontological Shifts:
From Individual Cognition to Relational Assemblages
Traditional educational models are
based on an individualist ontology, viewing the learner as a discrete cognitive
subject and knowledge as an entity to be internalised and transferred. Within
this framework, technologies are regarded as neutral tools that facilitate
learning without changing its essential character. In contrast, AI-mediated
environments challenge this assumption by actively participating in processes
such as idea generation, feedback provision, and evaluative judgment.
Sociomaterial and posthuman
perspectives conceptualise learning as relational, distributed, and emergent,
arising from interactions among learners, technologies, discourses, and
institutional structures (Fenwick et al., 2015; Barad, 2007). Within this framework,
AI systems function as epistemic actors that influence what is visible,
conceivable, and valued in learning environments (Knox, 2019). As a result, the
learner is repositioned as a component of a human–AI assemblage rather than the
exclusive site of cognition.
Epistemological
Implications: Knowledge, Authority, and Provisionality
AI systems produce outputs through
probabilistic modelling and extensive data training, resulting in responses
that are fluent but inherently uncertain and context-dependent. Nevertheless,
learners may perceive AI-generated content as authoritative because of its
speed, coherence, and institutional legitimacy (Williamson & Eynon, 2020).
This dynamic introduces significant epistemological challenges concerning the
production, validation, and trustworthiness of knowledge in AI-mediated
learning environments.
When AI systems deliver immediate
answers or suggestions, learners may conflate correctness with genuine
understanding or assume that epistemic responsibility resides with the system.
These dynamics risk privileging superficial performance over deep sense-making
and reflective engagement (Biesta, 2015; Selwyn, 2023). An interpretivist
perspective thus frames knowledge in AI contexts as situated and negotiated,
rather than objective or immutable.
Metacognition in Distributed Cognitive
Systems
Reframing
Metacognition in AI-Rich Environments
Metacognition is traditionally defined
as learners’ awareness and regulation of their cognitive processes,
encompassing planning, monitoring, and evaluation (Flavell, 1979). Although
these elements remain pertinent, AI-mediated learning environments complicate
metacognitive activity by externalising cognitive processes that were once
internal to the learner.
AI tools can scaffold metacognition by
modelling strategies, prompting reflection, or offering adaptive feedback
(Azevedo et al., 2017). However, these tools may also obscure cognitive effort,
creating illusions of competence as learners overestimate their understanding
due to the polished nature of AI-generated outputs (Dunlosky & Rawson,
2012). This tension necessitates a reconceptualisation of metacognition that
incorporates distributed and delegated cognition.
Metacognition as
Ontological Competence
In AI-mediated learning environments,
metacognition extends beyond monitoring one’s own thinking to include reflexive
awareness of where cognition occurs and how agency is distributed between
humans and machines. Learners must develop the capacity to recognise when
cognitive labour has been delegated to AI systems, when such delegation is
pedagogically productive, and when it may undermine learning or autonomy.
Framing metacognition as an
ontological competence aligns with recent scholarship on critical AI literacy,
which highlights the importance of learners understanding the limitations,
biases, and socio-technical implications of AI systems (Ng et al., 2021;
Bearman et al., 2023). This perspective positions metacognition as essential
for maintaining learner agency in increasingly automated educational
environments.
Interpretivist
Theoretical Alignment
This theoretical framework is grounded
in an interpretivist qualitative paradigm, which posits that reality is
socially constructed, contextually situated, and experienced through
meaning-making practices (Lincoln & Guba, 1985). Instead of viewing AI’s
educational impact as an objective phenomenon to be measured, the interpretivist
approach seeks to understand how learners and educators experience, interpret,
and negotiate AI-mediated learning.
Within this paradigm, metaphysical
assumptions about learning are explored as lived realities enacted through
classroom practices, institutional discourses, and daily interactions with AI
systems. Metacognition is similarly understood as a situated and interpretive
process, influenced by learners’ identities, cultural backgrounds, and prior
educational experiences. This approach is particularly appropriate for
inclusive and international education contexts, where concepts of independence,
support, and success are culturally mediated and contested.
Inclusive Classrooms
and International School Contexts
In inclusive classrooms, AI
technologies are often positioned as assistive tools intended to support
learners with diverse cognitive profiles. Although these applications offer
considerable potential for enhancing accessibility, they may inadvertently reinforce
deficit-oriented narratives if learners’ cognition is implicitly regarded as
incomplete without technological augmentation (Oliver, 2016). A metaphysically
informed perspective reframes AI as one participant within a broader learning
ecology, supporting diverse epistemologies without undermining learner agency.
International school contexts add
further complexity. AI systems frequently privilege dominant linguistic norms
and Western epistemologies, influencing which forms of knowledge are recognised
and legitimised (Andreotti et al., 2015). For learners with transnational
identities, AI may serve both as a resource for mobility and as a mechanism of
epistemic standardisation. Metacognitive awareness is therefore crucial for
enabling learners to critically interpret AI feedback and situate it within
their own cultural and educational trajectories.
Design Principles for
Metacognitively Responsible AI
Based on the preceding analysis, the
following design principles are recommended:
- Transparency of
Cognitive Roles – AI systems should make explicit how their outputs are generated
and what role they play in learning processes.
- Design for
Reflective Friction – AI should introduce moments of pause, questioning, or explanation
to support deliberation and sense-making.
- Support Without
Substitution – Scaffolds should be adjustable and fadeable to preserve learner
agency.
- Plurality of
Ways of Knowing – Systems should accommodate diverse cognitive styles and
representations, particularly for neurodiverse learners.
- Critical AI
Literacy by Design – Learners should be supported to question AI outputs and recognise
limitations and bias.
- Ethics as
Embedded Practice – Ethical considerations should be integrated into everyday
interactions rather than treated as external constraints.
Conclusion
This paper has demonstrated that
AI-mediated learning environments require a re-examination of the metaphysical
and metacognitive foundations of education. By situating AI within relational
ontologies of learning and reconceptualising metacognition as an ontological
competence, the proposed framework foregrounds learner agency, inclusion, and
ethical engagement. Anchored in an interpretivist paradigm, this analysis
prioritises learners' lived experiences and meaning-making practices, providing
theoretical and practical insights for inclusive and international education in
the context of intelligent technologies.
References
Andreotti, V. de O., Stein, S.,
Pashby, K., & Nicolson, M. (2015). Social cartographies as performative
devices in research on higher education. Higher Education Research &
Development, 35(1), 84–99. https://doi.org/10.1080/07294360.2015.1125857
Azevedo, R., Taub, M., & Mudrick,
N. V. (2017). Understanding and reasoning about real-time cognitive, affective,
and metacognitive processes to foster self-regulated learning with advanced
learning technologies. Handbook of Self-Regulation of Learning and
Performance, 254–270. https://doi.org/10.4324/9781315697048
Barad, K. (2007). Meeting the
universe halfway: Quantum physics and the entanglement of matter and meaning.
Duke University Press.
Bearman, M., Ajjawi, R., & Tai, J.
(2023). Beyond “AI literacy”: Why critical engagement with artificial
intelligence matters for education. Teaching in Higher Education.
Advance online publication. https://doi.org/10.1080/13562517.2023.2186127
Biesta, G. (2015). Good education
in an age of measurement: Ethics, politics, democracy. Routledge.
Dunlosky, J., & Rawson, K. A.
(2012). Overconfidence produces underachievement: Inaccurate self-evaluations
undermine students’ learning and retention. Learning and Instruction, 22(4),
271–280. https://doi.org/10.1016/j.learninstruc.2011.08.003
Fenwick, T., Edwards, R., &
Sawchuk, P. (2015). Emerging approaches to educational research: Tracing the
socio-material. Routledge.
Flavell, J. H. (1979). Metacognition
and cognitive monitoring: A new area of cognitive–developmental inquiry. American
Psychologist, 34(10), 906–911. https://doi.org/10.1037/0003-066X.34.10.906
Knox, J. (2019). What does the
“postdigital” mean for education? Three critical perspectives on the digital,
with implications for educational research and practice. Postdigital Science
and Education, 1(2), 357–370. https://doi.org/10.1007/s42438-019-00045-y
Lincoln, Y. S., & Guba, E. G.
(1985). Naturalistic inquiry. Sage.
Ng, D. T. K., Leung, J. K. L., Chu, S.
K. W., & Qiao, M. S. (2021). AI literacy: Definition, teaching, evaluation,
and ethical issues. Proceedings of the Association for Information Science
and Technology, 58(1), 504–509. https://doi.org/10.1002/pra2.487
Oliver, M. (2016). Accommodating
learners, contexts, and technologies: Reconsidering the ‘problem’ of
disability. Journal of Interactive Media in Education, 2016(1), Article
4. https://doi.org/10.5334/jime.399
Selwyn, N. (2019). Should robots
replace teachers? AI and the future of education. Polity Press.
Selwyn, N. (2023). Education and
technology: Key issues and debates (3rd ed.). Bloomsbury.
Williamson, B., & Eynon, R.
(2020). Historical threads, missing links, and future directions in AI in
education. Learning, Media and Technology, 45(3), 223–235. https://doi.org/10.1080/17439884.2020.1798995
Williamson, B., Eynon, R., &
Potter, J. (2020). Pandemic politics, pedagogies, and practices: Digital
technologies and distance education during the coronavirus emergency. Learning,
Media and Technology, 45(2), 107–114. https://doi.org/10.1080/17439884.2020.1761641
Comments
Post a Comment