Introduction: Large language models (LLMs) are increasingly employed as cognitive aids in research and professional inquiry, yet their fluent outputs are frequently regarded as authoritative knowledge. We contend that this practice signifies a fundamental epistemic misalignment. Methods/Approach: Building on Peirce’s theory of inquiry, Sellars’ concept of the space of reasons, Stanovich’s tripartite model of cognition, and knowledge-building theory, we develop a conceptual framework for analyzing epistemic agency in human–LLM collaboration. Results/Argument: We demonstrate that LLM outputs fail to satisfy the conditions for knowledge because they lack reflective regulation, resistance to revision, and normative commitment. While LLMs display strong autonomous and algorithmic abilities (e.g., pattern recognition and hypothesis development), reflective control remains a distinctly human function. This asymmetry supports a principled division of epistemic labour and motivates the concept of the Knowledge-Building Partner (KBP): an AI system designed to support inquiry without claiming epistemic authority. Discussion/Implications: We identify prompt-, system-, and model-level design requirements and introduce a triangulated framework for operationalizing epistemic agency through explainable AI, discourse analysis, and rational-thinking measures. These contributions collectively reposition LLM limitations as epistemic design challenges rather than technical issues.
2 February 2026 in Journal of Digital Life and Learning
This mixed-methods case study examines a redesigned 12-week teacher education course on Child Development in which weekly mid-week microlessons were embedded as core companions to lectures and readings. Each microlesson was designed in accordance with evidence-informed practices (e.g., spaced exposure, retrieval opportunities) and Mayer’s multimedia learning principles. Twenty-six teacher candidates completed 12 microlessons over the semester. Data sources included web server logs, embedded mini-surveys, an end-of-term survey, and semi-structured interviews. Learners engaged frequently with the microlessons (826 total visits; ≈2.6 views per lesson per student), with 41% of views occurring within 48 hours of release, indicating distributed engagement. Contrary to assumptions about mobile learning, 94% of accesses occurred on laptops or desktops, which students associated with focused academic work. Participants described the course as well-paced and manageable; the combination of shorter synchronous sessions, spaced microlessons, and low-stakes multiple-attempt quizzes supported attention, self-regulation, and retrieval practice. Students reported revisiting microlessons to clarify ideas, prepare assignments, and model designs for their own teaching. Findings suggest that well-designed microlessons can enhance teacher education by segmenting and distributing key content while fostering repeated, purposeful engagement. Limitations include the single-course context, small sample, and the absence of direct learning-outcome measures.
This study examined preservice teachers’ use of generative Artificial Intelligence (AI) chatbots for lesson planning during K-12 practicum placements. Using a convergent mixed-methods approach, we analyzed survey responses (n = 103; response rate = 12.88%) and conducted 4 semi-structured interviews to contextualize quantitative patterns. Among respondents, approximately three-quarters reported using AI chatbots for lesson-planning-related tasks, with ChatGPT being the most common tool. Exploratory analyses (χ² tests and a focused logistic model predicting frequent use (≥ half the time)) suggest that use frequency is driven more by perceived value and practical needs than by demographic factors. Interview data indicate that participants primarily used AI to save time and generate ideas, particularly under time pressure or when teaching less familiar topics, while noting risks related to curriculum alignment, accuracy, technical reliability, and over-reliance. Given the low response rate and likely self-selection, findings should be interpreted as indicative rather than fully generalizable. We conclude with program-level recommendations, including a minimal practicum policy toolkit (Appendix D) to support responsible and transparent AI use in teacher education.