Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis

Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 AMPLIFIER AND RISK: A THEORY-BUILDING EXTENSION OF THE COMMUNITY OF INQUIRY (COI) FOR STUDENT AI USE Sharlene Baksh Concordia University, Montreal (Canada) & The University of the West Indies - UWI Global Campus (Jamaica) ORCID: https://orcid.org/0009-0004-4629-1154 EMAIL: sharlene.bak[email protected]u ABSTRACT Purpose: This study extends the Community of Inquiry (CoI) framework by theorising two student-derived constructs, Cognitive Amplifier and Ethical Risk, to better capture how generative AI shapes learning in Caribbean higher education contexts. Methodology: Guided by a pragmatist, social-constructivist stance, this theory-building study conducted a secondary qualitative analysis using 68 open-ended narratives drawn from a larger mixed-methods study of 114 tertiary students. Deductive coding applied CoI categories, while inductive coding identified emergent themes. Teaching Presence was interpreted as a moderating factor. Findings: Students described AI as a Cognitive Amplifier, enriching meaning making by clarifying complex ideas, generating examples, and providing multimodal scaffolding. They also expressed strong Ethical Risk, including anxieties about misinformation, plagiarism, and policy ambiguity, which discouraged disclosure and weakened collaboration. Across both dynamics, Teaching Presence emerged as a key influence whether AI was experienced as supportive or stressful, suggesting its potential role as a moderating lever rather than a causal determinant. Value: The study positions Cognitive Amplifier and Ethical Risk as cross-cutting mechanisms rather than new presences, refining CoI to address AI-mediated inquiry. This theory-building contribution advances a more globally representative understanding by foregrounding Caribbean student perspectives, which are often underrepresented in AI in education research. Practical Implications: The extended framework suggests actionable strategies for higher education, including transparent institutional policies, embedded AI literacy curricula, integrity by design assessments and encouraging the responsible use of integration of multimodal AI scaffolding into teaching practice. Keywords: AI in education, Community of Inquiry (CoI), Cognitive Amplifier, Ethical Risk, Teaching Presence, Higher education, Caribbean, Global South Received: 14 september 2025 / Revised: 21 november 2025/ Accepted: 3 december 2025 DOI: https://doi.org/10.37497/rev.artif.intell.educ.v6ii.57 Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. AMPLIFICADOR E RISCO: UMA EXTENSÃO DE CONSTRUÇÃO TEÓRICA DO MODELO COMMUNITY OF INQUIRY (COI) PARA O USO DE IA POR ESTUDANTES RESUMO Propósito: Este estudo estende o framework Community of Inquiry (CoI) ao teorizar dois construtos derivados da experiência estudantil — Amplificador Cognitivo e Risco Ético — com o objetivo de compreender como a inteligência artificial generativa molda a aprendizagem em contextos do ensino superior no Caribe. Metodologia: Orientado por uma postura pragmatista e socioconstrutivista, este estudo de construção teórica realizou uma análise qualitativa secundária utilizando 68 narrativas abertas provenientes de uma pesquisa maior, de métodos mistos, com 114 estudantes do ensino superior. A codificação dedutiva aplicou as categorias do CoI, enquanto a codificação indutiva identificou temas emergentes. A Presença de Ensino foi interpretada como um fator moderador. Resultados: Os estudantes descreveram a IA como um Amplificador Cognitivo, enriquecendo a construção de sentido ao esclarecer ideias complexas, gerar exemplos e oferecer suportes multimodais. Também relataram forte percepção de Risco Ético, incluindo ansiedade sobre desinformação, plágio e ambiguidade das políticas institucionais — fatores que desencorajavam a divulgação do uso de IA e enfraqueciam a colaboração. Em ambas as dinâmicas, a Presença de Ensino emergiu como um elemento-chave para determinar se a IA era percebida como apoio ou como fonte de estresse, sugerindo seu papel potencial como alavanca moderadora e não como determinante causal. Contribuição: O estudo posiciona Amplificador Cognitivo e Risco Ético como mecanismos transversais, e não como novas “presenças”, refinando o CoI para abarcar processos investigativos mediados por IA. Essa contribuição teórica reforça uma compreensão mais globalmente representativa ao destacar perspectivas de estudantes caribenhos, frequentemente subrepresentadas na pesquisa sobre IA na educação. Implicações Práticas: O framework ampliado sugere estratégias acionáveis para o ensino superior, incluindo políticas institucionais transparentes, currículos de letramento em IA, avaliações estruturadas com foco em integridade e o incentivo ao uso responsável da IA multimodal integrada às práticas pedagógicas. Palavras-chave: IA na educação; Community of Inquiry (CoI); Amplificador Cognitivo; Risco Ético; Presença de Ensino; Ensino Superior; Caribe; Sul Global. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Introduction Generative artificial intelligence (AI) tools such as ChatGPT, Grammarly, and Quillbot are rapidly reshaping higher education. For students, these tools can clarify complex concepts, accelerate tasks, and scaffold independent study. In parallel, concerns about accuracy, academic integrity, and overreliance raise new forms of risk. Despite the accelerating integration of AI into academic life, existing theoretical models, most notably the Community of Inquiry (CoI) framework, were developed in pre-AI contexts and do not yet account for how students themselves experience both the benefits and anxieties of AI, particularly in under-resourced settings such as the Caribbean. This paper addresses that gap by extending the CoI framework with two constructs grounded in student narratives from a Caribbean higher education study. These constructs capture how AI is experienced simultaneously as an amplifier of inquiry and as a source of ethical tension. By theorising them within CoI, the study refines a widely used framework, foregrounds student perspectives from the Global South1, and offers practical implications for policy, pedagogy, and future research. This builds on calls from Southern theory (Connell, 2007) and postcolonial education scholarship (Tikly, 2019) to foreground knowledge from underrepresented regions in rethinking global frameworks. 1 In the Caribbean, regional analyses (Alexander, Diouf, & Wooding, 2025) also highlight how infrastructural inequities and policy gaps influence the adoption of emerging technologies. The paper proceeds in five steps. First, the literature review situates current debates on AI, learning, and integrity within the context of higher education. Second, the conceptual framework introduces the proposed extensions to CoI and positions them in relation to existing scholarship. Third, the methodology explains the secondary qualitative analysis of 68 student narratives. Fourth, the findings illustrate how students voiced both opportunities and risks in their accounts of AI use. Ultimately, the discussion highlights theoretical contributions, provides practical strategies for institutions, and outlines directions for future research. Guided by a pragmatist, social constructivist stance, the study positions these constructs as conceptual tools for refining CoI in AI mediated learning, rather than as new presences. Literature Review Research on generative AI in higher education has expanded rapidly, highlighting both its pedagogical potential and its ethical complexities. Existing studies examine how AI tools support meaning making, reshape collaboration, and challenge academic integrity. Simultaneously, the Community of Inquiry (CoI) framework remains a central model for analysing online and blended learning, yet it was developed before the rise of generative AI. This 1 In this paper, the term Global South refers to regions historically shaped by colonialism and structural inequities in resources, including digital infrastructure, whereas Global North refers broadly to industrialised regions (e.g., Europe, North America, parts of East Asia) that dominate knowledge production and policy discourse (Connell, 2007; Tikly, 2019). Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. review situates the study within current scholarship by examining three strands of literature most relevant to extending CoI: (1) AI as cognitive and multimodal scaffolding, (2) ethical risks and student anxieties, and (3) the evolving role of teaching presence in moderating AI use. AI as Cognitive and Multimodal Scaffolding Generative AI is increasingly recognised as more than a text-based assistant, with emerging scholarship highlighting its role in multimodal scaffolding that supports deeper meaning making. While topic-modelling analyses of over 4,000 publications show that text-to-text applications still dominate the field, modalities such as text-to-image and image-to-text remain comparatively underexplored (Heilala, Araya, & Hämäläinen, 2025). This gap is significant evidence given that multimodal AI expands accessibility and engagement by providing multiple entry points into learning (Arar et al., 2025). Recent case studies demonstrate how multimodal AI can be designed for personalised learning, combining text interaction, diagram interpretation, and image generation to scaffold complex problem solving (Chan, 2023). In teacher education, at the preservice level, teachers engaged with generative AI for multimodal composing, blending textual, visual, and design elements into their work, which shifted both their creative processes and their reflections on authorship (Khalil et al., 2023). Institutional pilots have also shown practical applications. Cooper and Tang (2024) reported that students used AI-generated images for brainstorming, creative writing, and visual research, though they required guidance to evaluate image quality and navigate ethical limitations. Within this same context, research cautions that not all AI-generated visuals equally support learning. Heilala, Araya & Hämäläinen (2025) distinguish between instructive, decorative, and distracting images, emphasising that visuals enhance comprehension only when they are directly aligned with the learning task. This echoes findings from Jin et al. (2025), Mohammadi et al. (2025), and Qian (2025), which showed that multimodal platforms deepen comprehension when paired with intentional human facilitation. When viewed together, the literature suggests that multimodal AI holds substantial promises for strengthening Cognitive Presence within the Community of Inquiry framework, but its effectiveness depends on thoughtful design, alignment with pedagogical goals, and the cultivation of AI literacy. Rather than treating AI’s multimodal capacity as inherently beneficial, the evidence highlights the need for educators to guide students in discerning when and how multimodal scaffolding contributes meaningfully to inquiry. Ethical Risks, Academic Integrity, and Student Anxiety Alongside AI’s pedagogical potential, a growing body of research points to the ethical ambiguities that shape how students engage with these tools. A systematic review in Technology in Society underscores that academic integrity in AI-rich environments is increasingly complex, with blurred boundaries between assistance and misconduct (Khalil et al., 2023). Large-scale surveys confirm this tension. Lund, Lee, Mannuru & Arutla (2025) found that students distinguished between “whole-AI writing,” widely judged as misconduct, and limited uses such as editing or idea Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. generation, which many perceived as acceptable. These differences reveal a persistent gap between institutional policies and students' intuitions about appropriate use. Broader sector research amplifies these concerns. Lund et al. (2025) found that students perceive clear distinctions between acceptable and unacceptable uses of generative AI, with strong anxieties about misconduct and uncertainty over institutional readiness. In the Caribbean, limited institutional guidance has compounded this uncertainty, with students reporting heightened anxiety about whether AI use is permissible at all (Bissessar, 2023; Roberts et al., 2024). Together, these studies validate the construct of Ethical Risk: a student-derived expression of anxiety linked to misinformation, plagiarism, and policy ambiguity that discourages disclosure and undermines collaborative trust. Research also highlights the broader effects of ethical ambiguity. Unclear boundaries do not only shape compliance; they also reshape behaviour, encouraging private or hidden use of AI rather than open, collaborative engagement. Lund, Lee, Mannuru & Arutla (2025) and Roberts et al. (2024) show that uncertainty about acceptable AI use often leads to anxiety, silence, or avoidance. Caribbean findings echo this pattern: students left to self-regulate without guidance become hesitant to share or collaborate (Bissessar, 2023; Roberts et al., 2024). Ethical Risk, therefore, functions not simply as an integrity issue but as a suppressor of Social Presence, weakening the openness and trust that CoI assumes as foundational to inquiry. Teaching Presence: Policy, Literacy, and Pedagogical Design Within CoI, Teaching Presence has long been recognised as the design and facilitation of learning processes (Anderson et al., 2001). Recent empirical studies continue to affirm its centrality. For instance, Vietnamese students described Teaching Presence as the most influential factor shaping engagement and satisfaction in virtual learning environments (Pham & Nguyen, 2025). Such findings reinforce the view that clarity, facilitation, and feedback remain decisive components of effective instructional design, even across diverse cultural and technological contexts. Biagini (2025) emphasises the need for curricula that blend technical competence with ethical reasoning, and recent global analyses confirm that higher education institutions are rapidly developing policies and guidelines to govern the responsible use of generative AI (Jin et al., 2024; An et al., 2025). Policy analyses point to the importance of culturally responsive frameworks to ensure that rules and expectations resonate across diverse student populations (Grab, 2025). At the classroom level, emerging research also shows how generative AI can be designed to support reflective practice and ethical reasoning; for example, ChatGPT and similar tools can promote metacognitive engagement when students are prompted to question, verify, and reflect on outputs (Hsiao & Tang, 2024). Such approaches align with calls to integrate verification and ethical reflection directly into the learning process. Synthesising across these strands, prior studies show that clear policies reduce ambiguity, AI literacy equips students to critically evaluate outputs, and integrity-by-design assessments align incentives with reflective practice (Roberts et al., 2024; Biagini, 2025; Jin et al., 2024). While these studies do not describe this dynamic as moderation, interpreted within CoI they suggest that Teaching Presence may function as a moderating lever, shaping Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. whether AI is experienced as an amplifier of inquiry or as a source of anxiety. In this framing, such practices reinforce Social Presence by establishing shared expectations and strengthen Cognitive Presence by embedding verification and reflection, ensuring that AI serves as a scaffold for meaningful learning rather than a threat to it. The literature confirms both the opportunities and tensions of AI in higher education. Generative AI offers new pathways for meaning making through multimodal scaffolding, yet unresolved issues of accuracy, misconduct, and policy ambiguity heighten ethical risk and suppress collaborative engagement. Across these dynamics, teaching presence consistently emerges as the decisive factor in shaping whether AI supports or undermines inquiry. However, existing frameworks, including CoI, were developed prior to the rise of generative AI and do not yet account for the dual role of AI as both amplifier and risk. This gap underscores the need for a refined theoretical model that integrates student-derived constructs, positioning teaching presence as the moderating device in AI-mediated learning. Conceptual Framework The Community of Inquiry (CoI) framework conceptualises meaningful learning as the interplay of three presences: Cognitive, Social, and Teaching (Garrison, Anderson, & Archer, 2000). While CoI remains robust, the emergence of generative AI introduces dynamics that extend beyond the original model. Students in the present study consistently described AI as both a partner in meaning making and a source of ethical tension. To capture these dual dynamics, two constructs are proposed: Cognitive Amplifier and Ethical Risk2. Figure 1 illustrates the original CoI framework, highlighting how meaningful learning emerges from the intersection of Cognitive, Social, and Teaching Presences. Figure 1. Original Community of Inquiry (CoI) framework, illustrating the interplay of Cognitive, Social, and Teaching Presences. Adapted from Garrison, Anderson, and Archer (2000). Re-created by the author. 2 2 Cognitive Amplifier refers to the ways AI enhances meaning making by clarifying concepts, generating examples, and providing multimodal scaffolding (e.g., visuals, diagrams). Ethical Risk refers to student anxieties about misinformation, plagiarism, and policy ambiguity. It does not denote actual misconduct or institutional rules, but the lived uncertainty that suppresses openness and collaboration. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Cognitive Amplifier refers to the ways in which AI enhances meaning making by clarifying concepts, generating examples, and offering multimodal scaffolding such as diagrams and visualisations. This construct extends CoI by specifying how AI augments the processes of exploration, integration, and resolution that underpin Cognitive Presence. Ethical Risk captures students’ anxieties around misinformation, plagiarism, policy ambiguity, and the erosion of independent thinking. These concerns suppress openness and collaboration, thereby constraining Social Presence. Importantly, these constructs are not proposed as new “presences” but as cross cutting, technology mediated mechanisms. Their influence is shaped by Teaching Presence, which students identified as the decisive factor determining whether AI functioned as a supportive amplifier or as a source of stress. This moderating role is interpreted here as a theoretical proposition for future testing, rather than a causal claim established by the present data. Positioning and Boundary Conditions To ensure clarity, the constructs are positioned against existing CoI elements, related extensions, and broader digital education literature (see Table 1). Cognitive Amplifier overlaps with but is not equivalent to Cognitive Presence. It specifies AI’s role as a mediating partner, not the student’s meaning making itself. Ethical Risk overlaps with but is not reducible to issues of academic integrity. It captures the lived anxiety and hesitancy that suppresses collaboration. Related proposals, such as Learning Presence or self-regulation, highlight metacognition and agency, while the constructs here foreground AI’s mediating role. In digital ethics, concerns such as misinformation and misconduct are well-documented, but Ethical Risk specifies how these translate into suppressed Social Presence within CoI. Together, the constructs can be understood as bidirectional continua: Amplification ↔ Attenuation of cognitive engagement, and Ethical Risk ↔ Ethical Assurance. Framing them in this way underscores that AI’s effects are contingent on pedagogical and institutional design. Table 1 summarises how the constructs are positioned in relation to CoI, related extensions, and digital education literatures. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Table 1. Positioning of Student-Derived Constructs within and beyond the Community of Inquiry (CoI). Developed by the author. Note: Table 1 positions the student-derived constructs of Cognitive Amplifier and Ethical Risk within the broader Community of Inquiry (CoI) framework. Core attributes and mechanisms were derived from prior CoI scholarship (Garrison, Anderson, & Archer, 2000; Anderson et al., 2001) as well as recent studies of AImediated learning. Evidence for Cognitive Amplifier draws on multimodal scaffolding research (Heilala, Araya, & Hämäläinen, 2025; Mohammadi et al. 2025; Qian 2025) and qualitative student accounts from Caribbean contexts (Baksh, 2025). Evidence for Ethical Risk reflects studies documenting student anxiety, policy ambiguity, and reluctance to disclose AI use (Bissessar, 2023; Roberts et al., 2024; Lund et al., 2025). Construct Definition Core Attributes Boundaries (NotThis) Evidence Anchors Consequences Instructional Implications Related Proposition(s) Cognitive Amplifier AI as a partner that enriches student meaning making within inquiry processes Clarifications, concrete examples, multimodal scaffolding (text, visuals, diagrams) Not identical to Cognitive Presence; not ease of use or performance expectancy Student quotes (R7, R9, R14); multimodal AI studies (Heilala et al., 2025) Deeper integration of knowledge; enhanced selfefficacy Encourage multimodal AI use; integrate verification and reflective activities P1, P4 Ethical Risk Student anxiety about misinformation, plagiarism, and unclear policies Ambivalence; fear of misconduct; reluctance to disclose use Not equivalent to actual misconduct; not institutional rules themselves Student quotes (R19, R27, R41); integrity/anxiety studies (Bissessar, 2023; Roberts et al., 2024) Suppressed Social Presence; hidden use; heightened anxiety Publish transparent AI policies; normalise disclosure; embed ethics in teaching P2, P3 Teaching Presence (as moderating lever) Instructional design and facilitation of AImediated learning shape whether AI is experienced as support or stress Policy clarity; AI literacy scaffolding; integrity-bydesign assessments Not a new presence; remains within CoI design or facilitation Student quotes (R12, R21, R33, R45); teaching presence research (Anderson et al., 2001; Biagini, 2025) Amplifies benefits of Cognitive Amplifier; attenuates harms of Ethical Risk Redesign assessments; embed AI literacy; provide explicit policy guidance P3, P4, P6 Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. While these studies do not use CoI terminology, they describe concealed or anxious engagement with AI. In this paper, these patterns are theorised as suppressed Social Presence, hidden use, and heightened anxiety. The moderating role of Teaching Presence is supported by research on instructional design and AI literacy practices (Biagini, 2025; Jin et al., 2024). Collectively, the table demonstrates boundary conditions, i.e., the constructs are not new presences, nor rebranded versions of existing categories, but cross-cutting mechanisms that operate transversally across CoI. Their positioning clarifies antecedents, mechanisms, and consequences, making them amenable to further theoretical refinement and empirical testing. As shown in Figure 2, the extended framework proposed in this study illustrates how Cognitive Amplifier and Ethical Risk act as cross-cutting mechanisms across the presences, with Teaching Presence moderating their influence. Figure 2. Extended CoI framework with student-derived constructs. Cognitive Amplifier and Ethical Risk act as cross-cutting mechanisms across presences, while Teaching Presence moderates their influence. Developed from Garrison et al. (2000) with new extensions proposed in this study. Developed by the author. Propositions: Extending the Community of Inquiry Two student-derived constructs, Cognitive Amplifier and Ethical Risk, extend the Community of Inquiry (CoI) framework by specifying how generative AI reshapes learning presences. The following propositions articulate these relationships, grounded in both prior scholarship and student narratives. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Ethical Considerations This secondary qualitative analysis was conducted using 68 open-ended narrative responses drawn from a previously approved mixed-methods study of tertiary students in the Caribbean (Baksh, 2025). The original study received institutional ethical approval, and participants were informed that they would be asked about their use of digital and AI tools for the study and assessment. For the present paper, no new data were collected, and no participants were re-contacted. All analysis was carried out on anonymised responses, consistent with the scope of the original ethics approval. Findings: Evidence for Inductive Extensions The secondary qualitative analysis of 68 narrative responses revealed two constructs that extend the Community of Inquiry (CoI) framework: Cognitive Amplifier and Ethical Risk. These constructs, voiced directly by students, illustrate how generative AI reshapes learning presence, with Teaching Presence consistently emerging as the moderating force. Cognitive Amplifier Students consistently portrayed AI as a Cognitive Amplifier that enriched the phases of exploration, integration, and resolution central to Cognitive Presence (Garrison et al., 2000). Many described how AI clarified complex concepts and allowed them to redirect time toward deeper engagement: “It allowed me to spend more time reading and examining complex information and data, rather than searching for it, improving my analytical skills and the quality of work I submit” (R14). Another explained: “AI tools can articulate complex information in a more digestible and memorable way, depending on the prompt used” (R9). Students also highlighted multimodality as a distinctive benefit: “I used AI to generate pictures to explain abstract ideas, and that made the content clearer” (R7). Students attributed different amplifier effects to specific tools. ChatGPT was often described as a “conversational explainer”, Grammarly as a writing assistant, and Quillbot as a paraphrasing aid. Copilot and Gemini were less commonly cited but noted for coding or alternative explanations. These tool-specific functions collectively shaped the broader construct of Cognitive Amplifier. In tandem, not all accounts were unambiguously positive. Some students noted that AI could confuse rather than clarify, “Sometimes AI gave me the wrong explanation, and I felt more lost” (R28). This ambivalence suggests that, while AI-enriched Cognitive Presence, for many, its contribution was contingent on accuracy, prompting, and verification. These excerpts show how students directly described AI, clarifying concepts, providing examples, and offering multimodal scaffolding. The analysis interprets these accounts as evidence of a broader construct, Cognitive Amplifier, which extends Cognitive Presence beyond text-based support to include multimodal meaning making. Importantly, Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. this emphasis on multimodality was explicitly reported by students themselves (e.g., R7), not only inferred from prior scholarship. Ethical Risk Alongside these benefits, students expressed significant ambivalence about AI use, captured in the construct of Ethical Risk. Concerns focused on misinformation, plagiarism, and fear of crossing unclear institutional boundaries. As one student explained, “AI tools … can sometimes generate false or fabricated sources … it’s important to doublecheck the output” (R27). Ethical Risk also varies by tool. ChatGPT raised concerns about misinformation and fabricated citations, Quillbot triggered anxieties about plagiarism detection, and Grammarly, while widely used, was sometimes viewed as a grey area where policy clarity was lacking. Such distinctions suggest that Ethical Risk is not uniform but mediated by the perceived institutional legitimacy and function of different AI tools. Other students emphasised anxiety, “I feel more anxious submitting work because of the fear that comes with using AI for assignments” (R19). One student admitted, “I think it can diminish your chances to be able to think for yourself” (R41). Yet not all students rejected AI. Some acknowledged its potential while simultaneously expressing fear,“I don’t think AI is all bad, but I am afraid to rely on it too much” (R46). This mixture of recognition and caution illustrates how Ethical Risk does not eliminate use but suppresses openness and disclosure, thereby weakening Social Presence. These student accounts highlight anxiety, uncertainty, and mistrust as recurring experiences. While the narratives provide direct evidence of hesitation and disclosure reluctance, the analysis interprets this pattern as the construct of Ethical Risk, a suppressor of Social Presence within the CoI framework. While these risks were described broadly, it is possible that perceptions may also vary by discipline or academic level. For example, students in writingintensive programmes may experience higher anxiety around paraphrasing tools than those in technical fields. This study does not disaggregate responses by subgroup, but such variation represents an important line of inquiry for future work. Teaching Presence as Moderator Across both dynamics, students consistently pointed to Teaching Presence as an important influence on whether AI was experienced as empowering or anxiety-inducing. Some described how clear instructional guidance amplified AI’s benefits, “Honestly, it has broken down certain topics for me to understand and even given a step-bystep process on how to solve it” (R12). Others emphasised that clarity about acceptable use reduced anxiety, “The lecturer’s guidance made me confident that my use of AI was acceptable” (R21). By contrast, in the absence of such guidance, anxiety was heightened, “I feel anxious submitting work because I don’t know if using AI is considered wrong” (R33). Another student added, “AI helped me, but only because my teacher explained when it was allowed” (R45). Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Students’ reflections demonstrate how instructional clarity and policy guidance directly shaped their confidence and openness in using AI. Building on these observations, the analysis interprets Teaching Presence as the moderating lever that determines whether AI is experienced as a Cognitive Amplifier or as a source of Ethical Risk. Table 5. Exemplar Quotes by Theme: Evidence, Ambivalence and Inference. Developed by the author. Theme Illustrative Quote Insight Cognitive Amplifier “AI allowed me to spend more time analysing instead of searching for data” (R14). Evidence “AI’s diagrams made abstract concepts clearer” (R7). Evidence “AI tools can articulate complex information in a more digestible way” (R9). Evidence “Sometimes AI gave me the wrong explanation, and I felt more lost” (R28). Ambivalent Ethical Risk “AI tools sometimes generate false or fabricated sources — I double-checked everything” (R27). Evidence “I feel more anxious submitting work because I don’t know if AI is considered wrong” (R19). Evidence “I think it can diminish your chances to be able to think for yourself” (R41). Evidence “I don’t think AI is all bad, but I am afraid to rely on it too much” (R46). Ambivalent Teaching Presence “The lecturer’s guidance made me confident that my use of AI was acceptable” (R21). Evidence “Honestly, it has broken down certain topics for me to understand…” (R12). Evidence “I feel anxious submitting work because I don’t know if using AI is considered wrong” (R33). Evidence “AI helped me, but only because my teacher explained when it was allowed” (R45). Inference Note. Quotes are categorised to distinguish between (a) Evidence, where student accounts directly supported the theme; (b) Ambivalent, where students expressed mixed or contradictory views, and (c) Inference, where the researcher’s interpretation connects student accounts to broader constructs such as moderation. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Discussion From a theory-building, pragmatist perspective, the study extends the Community of Inquiry (CoI) framework by theorising two student-derived constructs, Cognitive Amplifier and Ethical Risk, that illuminate how generative AI mediates learning in higher education. These constructs are advanced not as new “presences” but as conceptual tools, offered to refine CoI in contexts where AI reshapes meaning making and ethical risk taking. Their articulation reflects the pragmatic emphasis on generating propositions that are useful for guiding both future research and educational practice. These propositions are offered as conceptual tools rather than asserting definitive causal claims. Rather than treating these constructs as new “presences,” the analysis positions them as cross cutting mechanisms that operate across Cognitive, Social, and Teaching Presences. Their emergence from student narratives underscores the need to refine CoI for contexts shaped by AI technologies. These insights are offered as theoretically informed propositions rather than empirical proofs, contributing to the ongoing evolution of CoI as a framework for understanding digitally mediated inquiry. Cognitive Amplifier: Expanding Cognitive Presence Students frequently described AI as a partner in meaning-making. It clarified complex ideas, provided concrete examples, and offered multimodal scaffolding such as diagrams and visuals. These accounts suggest that AI can function as a Cognitive Amplifier, enriching the exploration, integration, and resolution phases that define Cognitive Presence. Concurrently, some students reported confusion or misinformation, indicating that the contribution of AI is contingent on prompting, accuracy, and verification. Cognitive Amplifier should therefore be understood as a potential mediator of meaning making, not a uniform or guaranteed enhancement. These patterns illustrate the Amplification ↔ Attenuation continuum: while many students experienced AI as enriching inquiry, others reported confusion or misinformation, indicating that its impact is contingent on accuracy, prompting, and verification. Ethical Risk: Suppressing Social Presence Alongside perceived benefits, students voiced significant anxieties about misinformation, plagiarism, and policy ambiguity. Prior studies confirm that such ambiguity promotes anxiety and reluctance to disclose AI use (Bissessar, 2023; Roberts et al., 2024). While these studies do not frame their findings in CoI terms, interpreted within CoI they suggest the construct of Ethical Risk, in which uncertainty suppresses openness and trust, thereby constraining Social Presence. Unlike academic misconduct itself, Ethical Risk reflects the lived experience of uncertainty that discourages disclosure and collaboration. Recognising this dynamic refines CoI by showing that ethical ambiguity functions not only as a compliance concern but also as a suppressor of presence within learning communities. This reflects the Risk ↔ Assurance continuum. In contexts of policy ambiguity, students gravitated toward concealment and anxiety, whereas clear guidance provided ethical assurance that sustained trust and collaboration. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Teaching Presence as a Moderating Lever Across both dynamics, Teaching Presence consistently appeared as an influence shaping whether AI was experienced as supportive or stressful. Students highlighted that clear instructional guidance, transparent policies, and integrity by design assessments reduced Ethical Risk while amplifying cognitive benefits (Roberts et al., 2024; Biagini, 2025). While these studies and student accounts do not frame their findings in terms of moderation, interpreted within CoI they suggest that Teaching Presence may act as a moderating lever, influencing the balance between amplification and risk. This interpretation is conceptually compelling but not empirically tested. It is therefore advanced here as a theoretical proposition that warrants future examination. Theoretical and Practical Contributions Theoretically, the study demonstrates that the Community of Inquiry (CoI) remains a robust framework but requires refinement to capture AI-mediated learning. Cognitive Amplifier enriches CoI’s explanatory power by specifying AI’s role in multimodal meaning making. Ethical Risk sharpens the model by identifying how anxiety and policy ambiguity can constrain collaboration. Teaching Presence is reconceptualised to include governance of AI use through clarity, scaffolding, and assessment design. Practically, these refinements point toward actionable strategies that operate across multiple levels of higher education. At the institutional level, universities can publish transparent policies on AI use, embed AI literacy within curricula, and create recognition systems that reward ethical and creative engagement with AI. At the faculty and departmental level, programs can redesign assessments to emphasise process and originality, provide professional development on AI literacy, and encourage communities of practice to share effective pedagogical designs. At the instructor level, Teaching Presence can be enhanced by clarifying classroom expectations, integrating multimodal scaffolding into teaching, and modelling responsible prompting and evaluation. Finally, at the student level, practices can encourage open disclosure of AI use, provide opportunities to develop reflective strategies for evaluating AI outputs, and promote peer dialogue to establish shared norms of integrity and collaboration. Contextually, by foregrounding Caribbean student voices, the analysis underscores how structural inequities and policy gaps shape AI adoption in ways that are underexplored in Global North dominated research. This contribution resonates with broader Global South scholarship advocating for epistemic justice in digital education (Connell, 2007; Tikly, 2019). By situating Cognitive Amplifier and Ethical Risk in the lived experiences of Caribbean tertiary students, the study aligns with regional analyses (Alexander, Diouf, & Wooding, 2025) that call for culturally grounded approaches to AI governance and pedagogy. Limitations and Future Research As a secondary qualitative analysis, the study carries several limitations that should be acknowledged. First, the coding was conducted by a single researcher, which may limit intersubjective reliability. While trustworthiness was Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. enhanced through memo writing, peer debriefing, and an audit trail, no intercoder coefficient was calculated. Second, although the dataset included variation across disciplines and academic levels, the present study did not undertake systematic subgroup analysis. The findings should therefore be interpreted as cross-cutting patterns rather than discipline-specific insights. Third, the narratives were drawn from Caribbean tertiary students, whose experiences are shaped by regional inequities and policy gaps. As such, findings may not generalise globally, underscoring the importance of comparative research across Global North and Global South contexts. Finally, the study was designed as theory building rather than hypothesis testing. The constructs of Cognitive Amplifier and Ethical Risk should therefore be treated as conceptual tools for further refinement, rather than empirically verified categories. Future research can extend this work through a staged agenda. Expert content validation (e.g., Delphi panels) could refine construct boundaries and evaluate their applicability across diverse contexts. Quantitative testing using exploratory and confirmatory factor analysis (EFA/CFA) and structural equation modelling (SEM) would enable empirical assessment of the proposed relationships, particularly the moderating role of Teaching Presence. Recent CoI-based SEM research has shown that Teaching and Cognitive Presences significantly predict satisfaction and persistence, with moderating effects of learner characteristics such as gender and prior online experience (Paris, Lakhal, & Mukamurera, 2025). Comparative studies across regions and disciplines, including recent analyses of Teaching Presence in Vietnamese higher education (Pham & Nguyen, 2025), would further strengthen the generalisability of Amplifier and Risk dynamics and contribute to a more globally representative theorisation of AI in education. Conclusion This paper has extended the Community of Inquiry (CoI) framework to account for the dynamics of generative AI in higher education. Drawing on qualitative narratives from Caribbean tertiary students, two constructs were theorised. Cognitive Amplifier, which highlights how AI can enrich Cognitive Presence through clarifications, examples, and multimodal scaffolding; and Ethical Risk, which captures anxieties about misinformation, plagiarism, and policy ambiguity that can suppress Social Presence. Across both, Teaching Presence consistently appeared as a moderating influence on whether AI was experienced as supportive or uncertain, interpreted within CoI, this role can be understood as a potential moderating factor. The study makes three contributions. Theoretically, it refines CoI by positioning Cognitive Amplifier and Ethical Risk as cross cutting mechanisms rather than additional presences, clarifying how AI mediates inquiry processes. Practically, it identifies strategies that institutions can adopt, including transparent AI policies, embedded AI literacy, integrity-by-design assessments, and the use of multimodal scaffolding to deepen meaning making. Contextually, it amplifies Global South perspectives, illustrating how structural inequities and policy gaps shape AI adoption in ways often overlooked in Global North dominated research. In parallel, subgroup variation across disciplines and academic levels was not analysed here, representing an important limitation and future direction for testing the generalisability of Amplifier and Risk dynamics. Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. These insights should be understood as a contribution to theory building rather than a definitive empirical test. Framing these dynamics as Amplification ↔ Attenuation and Risk ↔ Assurance highlights that AI’s educational role is not fixed, rather, it depends on how institutional design and teaching presence shape its use. Future research can build on this foundation through expert content validation (e.g., Delphi panels) to sharpen construct boundaries, followed by quantitative testing using methods such as exploratory and confirmatory factor analysis (EFA/CFA) and structural equation modelling (SEM). Comparative studies across Global South and Global North institutions would further strengthen the generalisability of findings and advance a more globally representative theorisation of AI in education. In closing, the extended CoI framework developed here offers both conceptual refinement and actionable guidance. It underscores that AI is neither inherently an amplifier nor a risk. Instead, it is a mediating force whose educational value depends on how teaching presence, institutional design, and cultural context shape its use. Conflict of Interest Statement The author declares that there are no conflicts of interest with respect to the research, authorship, and/or publication of this article. References Alexander, D., Døhl Diouf, L., & Wooding, C. (2025). Artificial intelligence readiness in the Caribbean: An exploratory review. United Nations Economic Commission for Latin America and the Caribbean (ECLAC). https://caribbean.un.org/en/295561-artificial-intelligence-readiness-caribbean-exploratory-review Anderson, T., Rourke, L., Garrison, D. R., & Archer, W. (2001). Assessing teaching presence in a computer conferencing context. Journal of Asynchronous Learning Networks, 5(2), 1–17. https://doi.org/10.24059/olj.v5i2.1875 An, Y., Yu, J. H., & James, S. (2025). Investigating the higher education institutions’ guidelines and policies regarding the use of generative AI in teaching, learning, research, and administration. International Journal of Educational Technology in Higher Education, 22, Article 10. https://doi.org/10.1186/s41239-025-00507-3 Arar, K. H., Özen, H., Polat, G., & Turan, S. (2025). Artificial intelligence, generative artificial intelligence and research integrity: A hybrid systemic review. Smart Learning Environments, 12, Article 44. https://doi.org/10.1186/s40561025-00403-3 Baksh, S. (2025). Investigating Students’ AI Usage Through the Community of Inquiry (CoI) Framework: A MixedMethod Study in the Caribbean Region. SDG Studies, 3(1), 36. https://doi.org/10.37497/sdgs.v6istudies.36 Biagini, G. (2025). Towards an AI-literate future: A systematic literature review exploring education, ethics, and applications. International Journal of Artificial Intelligence in Education. Advance online publication. https://doi.org/10.1007/s40593-025-00466-w Bissessar, C. (2023). To use or not to use ChatGPT and assistive artificial intelligence tools in higher education institutions? The modern-day conundrum—students' and faculty's perspectives. Equity in Education & Society. https://doi.org/10.1177/27526461231215083 Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Braun, V., & Clarke, V. (2021). To saturate or not to saturate? Questioning data saturation as a useful concept for thematic analysis and sample-size rationales. Qualitative Research in Sport, Exercise and Health, 13(2), 201–216. https://doi.org/10.1080/2159676X.2019.1704846 Chan, C. K. Y. (2023). Is AI changing the rules of academic misconduct? An in-depth look at students’ perceptions of “AI-giarism.” arXiv. https://arxiv.org/abs/2306.03358 Chatfield, S. L. (2020). Recommendations for secondary analysis of qualitative data. The Qualitative Report, 25(3), 833–842. https://nsuworks.nova.edu/tqr/vol25/iss3/17/ Cooper, G., & Tang, K.-S. (2024). Pixels and pedagogy: Examining science education imagery by generative artificial intelligence. Journal of Science Education and Technology, 33, 556–568. https://doi.org/10.1007/s10956-02410104-0 Connell, R. (2007). Southern theory: The global dynamics of knowledge in social science. Polity Press. https://doi.org/10.4324/9781003117346 Grab, O., Balli, S. J., & Owusu, E. (2025). Teaching for Equity: An Exploration of AI’s Role in Culturally Responsive Teaching in Higher Education Settings. Innovative Higher Education. https://doi.org/10.1007/s10755-025-098014 Haidar, H., Suryoputro, G., & Safi’i, I. (2025). Impact of the integration of metacognitive prompts by generative AI in collaborative and individual learning in improving writing skills and metacognitive awareness. International Journal of Learning, Teaching and Educational Research, 24(6), 232-250. https://doi.org/10.26803/ijlter.24.6.11 Heilala, V., Araya, R., & Hämäläinen, R. (2025, March). Beyond text-to-text: An overview of multimodal and generative artificial intelligence for education using topic modeling. In Proceedings of the 40th ACM/SIGAPP Symposium on Applied Computing (SAC ’25) (pp. 54 - 63). ACM. https://doi.org/10.1145/3672608.3707764 Hsiao, H.-C., & Tang, Y. (2024). Beyond acceptance: An empirical investigation of technological, ethical, social, and individual determinants of generative AI–supported learning in higher education. Education and Information Technologies, 29(7), 9231–9254. https://doi.org/10.1007/s10639-024-13263-0 Jin, L., Yan, L., Echeverria, V., Martinez-Maldonado, R., & Gašević, D. (2024). Generative AI in higher education: A global perspective of institutional adoption policies and guidelines. Computers and Education: Artificial Intelligence, 6, 100228. [Preprint]. arXiv. https://arxiv.org/abs/2405.11800 Khalil, M., Prinsloo, P., & Slade, S. (2023). Artificial intelligence and academic integrity: A systematic review. International Journal for Educational Integrity, 19(1), 3. https://doi.org/10.1007/s40979-023-00129-0 Lund, B. D., Lee, T. H., Mannuru, N. R., & Arutla, N. (2025). AI and academic integrity: Exploring student perceptions and implications for higher education. Journal of Academic Ethics, 23(6), 1545–1565. https://doi.org/10.1007/s10805-025-09613-3 Mohammadi, M., Tajik, E., Martínez-Maldonado, R., Sadiq, S., Tomaszewski, W., & Khosravi, H. (2025). Artificial intelligence in multimodal learning analytics: A systematic literature review. Computers & Education: Artificial Intelligence, 8, 100426. https://doi.org/10.1016/j.caeai.2025.100426 Paris, J., Lakhal, S., & Mukamurera, J. (2025). Relationships between presence, satisfaction, and persistence in higher education online courses: Assessing the moderating effects of gender, age, and online experience using PLS-MGA. Educational Technology & Society, 28(1), 178–193. https://doi.org/10.30191/ETS.202501_28(1).RP10 Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Pham, A. T., & Nguyen, T. B. (2025). Students’ perceptions of teaching presence in the virtual learning environment: A Community of Inquiry case study in Vietnam. Acta Psychologica, 250, 104017. https://www.sciencedirect.com/science/article/pii/S0001691825011230 Qian, Y. (2025). Pedagogical applications of generative AI in higher education: A systematic review of the field. TechTrends. https://doi.org/10.1007/s11528-025-01100-1 Qu, H., Tan, H., & Wang, L. (2024). Disciplinary differences in undergraduate students’ engagement with generative artificial intelligence. Smart Learning Environments, 11(1), Article 17. https://doi.org/10.1186/s40561-024-003416 Roberts, L. N., Solomon, F. N., & Cummings, R. (2024). Navigating the AI landscape in Caribbean education. In AI in Education (pp. 268–285). IGI Global. https://doi.org/10.4018/979-8-3693-2728-9.ch012 Roberts, L. N. (2025). Navigating new frontiers: AI in assessments from the perspectives of teachers and students. In Effective instructional design informed by AI. IGI Global. https://www.igi-global.com/chapter/navigating-newfrontiers/369672 Tikly, L. (2019). Education for sustainable development in the postcolonial world: Towards a transformative agenda for Africa. Routledge. https://doi.org/10.4324/9781315211343 Wang, X., Zainuddin, Z., & Chin, H. L. (2025). Generative artificial intelligence in pedagogical practices: A systematic review of empirical studies (2022–2024). Cogent Education, 12(1), Article 2485499. https://doi.org/10.1080/2331186X.2025.2485499 Yan, L., Martinez-Maldonado, R., Jin, Y., Echeverria, V., Milesi, M., Fan, J., Zhao, L., Alfredo, R., Li, X., & Gašević, D. (2024). The effects of generative AI agents and scaffolding on enhancing students’ comprehension of visual learning analytics [Preprint]. arXiv. https://doi.org/10.48550/arXiv.2409.11645 Yang, X., Liu, X., & Gao, Y. (2025). The impact of generative AI on students’ learning: A study of learning satisfaction, self-efficacy and learning outcomes. Educational Technology Research and Development. Advance online publication. https://doi.org/10.1007/s11423-025-10540-8 Review of Artificial Intelligence in Education, Florida (EUA), V. 6, 01-25, e057, 2025 Baksh, S. (2025). Amplifier and Risk: Extending the Community of Inquiry (CoI) for Student AI Use: A Secondary Qualitative Analysis. Appendix 1 Adapted from Baksh, S. (2025). Investigating Students’ AI Usage Through the Community of Inquiry (CoI) Framework: A MixedMethod Study in the Caribbean Region. SDGs Studies Review, 6(studies), e036. https://doi.org/10.37497/sdgs.v6istudies.36 Summary Table of Qualitative Themes and Frequencies Theme Key Codes Frequency (total code hits) AI as a Cognitive and Performance Amplifier Cognitive Presence, Performance Expectancy, Teaching Presence 60, 43, 11 Effort, Access, and the Student Experience Effort Expectancy, Infrastructure/Access, Social Influence 40, 8, 5 Navigating Trust, Accuracy, and Ethics Ethical Concerns, Trust in AI 13, 12 Social and Emotional Dimensions of AI Use Social Presence, Emotional Impact 10, 3 Adaptive and Personalised Learning Paths Personalised Learning, Metacognitive Change 10, 10 Appendix 2 Adapted from Baksh, S. (2025). Investigating Students’ AI Usage Through the Community of Inquiry (CoI) Framework: A MixedMethod Study in the Caribbean Region. SDGs Studies Review, 6(studies), e036. https://doi.org/10.37497/sdgs.v6istudies.36 Integrated Joint Display Showing Alignment of Quantitative Variables (CoI, UTAUT) with Qualitative Themes and Codes Construct Quantitative Evidence Qualitative Evidence Cognitive Presence High means for AI clarifying concepts and connecting ideas (M = 4.24–4.17); strong correlation with perceived learning outcomes (ρ = 0.73, p < .001). Students describe AI as a “cognitive amplifier” aiding understanding and reflection (60 code hits). Performance Expectancy AI is seen as academically useful (M = 4.07); frequency of use correlated with usefulness (ρ = 0.51, p < .001). AI enhances performance, efficiency, and productivity; boosts revision and quality of work (43 hits). Teaching Presence Moderate mean score (M = 3.96); significant predictor of perceived usefulness (β = 0.437, R² = 0.261, p < .001). AI used as “digital tutor,” offering step-by-step guidance (11 hits). Effort Expectancy High ease-of-use ratings across levels (M = 4.08); no significant difference by academic level (U = 4306, p = .842). Most found AI tools intuitive and timesaving; some noted access limitations (40 hits). Social Presence / Collaboration Lower agreement that AI supports peer interaction (M = 3.22). Limited references to AI-enabled collaboration (10 hits); some mention improved expression. Ethical Concerns / Trust Moderate concern about ethical implications (M = 3.59). Concerns about plagiarism, misinformation, and appropriate use (13 hits). Personalised/Adaptive Learning Not directly measured in survey items. AI is viewed as a personalised study partner that adjusts to student needs and supports metacognitive development and learner autonomy (10 hits each: personalised, metacognitive). Insights extend beyond CoI/UTAUT.