Designing Microlearning Modules on AI for University Teachers in Vocational Teacher Education: A Design-Based Research Study

Synthesis: University lecturers face growing pressure to develop AI literacy, yet targeted professional development remains scarce. Cechovsky (2026) reports a design-based research project that developed and iteratively refined online microlearning modules on AI for lecturers in vocational teacher education. Two modules — AI in supervising bachelor's and master's theses and AI in competence-based Assessment — were built on a definition of microlearning and on cognitive load theory, then tested in two online group discussions held in October and December 2025. Participants valued the modular, on-demand format and its guided graphic structures, but flagged technical barriers (locating modules in OneNote, navigating Genially) and uncertainty about platform choice. From these converging theoretical and empirical strands the study derived five design principles: clear structure and orientation, guidance and navigation support, segmentation and modularization, Multimodal AI presentation, and learner activation. The author concludes that microlearning is promising for AI-related educational development in higher education, while cautioning that a small sample and the researchers' close involvement limit generalization.

Key Findings

1. University lecturers lack targeted AI support even as demand rises: a 2024 survey of 236 Austrian vocational teacher-training students found over 80% wanted to learn more about applying AI in teaching.
2. Two microlearning modules targeted core lecturer tasks — supporting scientific writing and supervising bachelor's and master's theses, plus competence-based assessment — covering pedagogical, legal, and ethical aspects.
3. Participants valued the modular, on-demand format, clearly delimited topic units, decision trees, and case examples, viewing these as reducing their cognitive burden when learning about AI.
4. Technical design caused friction: lecturers struggled to access modules in OneNote, navigate backwards within Genially, and enable audio for embedded videos, increasing extraneous cognitive load.
5. Participants judged OneNote too complex and proposed card-based alternatives — Padlet, Moodle, Taskcard, and Trello — singling out the flexibility of card-based platforms for delivering the modules.
6. Five design principles were derived: clear structure and orientation, guidance and navigation support, segmentation and modularization, multimodal presentation, and learner activation.
7. Limitations — a small sample, the researchers' close involvement with the field, and partly informal discussions — restrict generalization; future work targets vocational teacher-training students and quantitative evaluation.

Microlearning, cognitive load, and the AI literacy gap

University lecturers face mounting pressure to teach with and about AI, yet the support available to them lags behind the demand. The paper cites surveys that show the gap plainly: 1,558 university students across Germany, the US, and the UK reported high interest in AI but relatively low AI literacy, and a 2024 survey of 236 Austrian vocational teacher-training students found that over 80% wanted to learn more about applying AI in teaching, even as many were unsure about the rules governing AI in theses, assignments, and examinations. A UNESCO survey from September 2025 found that almost two-thirds of higher education institutions have AI guidance or are developing it, but more than half of respondents were uncertain about its educational implications. The AI Act (Regulation (EU) 2024/1689) sharpens the obligation by requiring that staff using AI systems be trained to do so competently and responsibly. Microlearning — short, media-supported units usable at any time — is offered as a flexible response, grounded in cognitive load theory and aimed at minimizing extraneous load.

Two evaluation cycles with vocational teacher educators

The project followed a four-phase design-based research cycle of preliminary examination, prototype development, assessment, and final solution. Starting in 2024, the team built two prototypes around core lecturer tasks: supporting students' scientific writing and thesis supervision, and AI in competence-based Assessment. Both deliberately addressed legal and ethical dimensions rather than the explanation of specific tools. The first online group discussion, held in October 2025, involved three university lecturers and focused on their experience of the modules; the second, in December 2025, involved four lecturers who actively engaged with digital media and concentrated on platform selection and unresolved design questions. Both sessions ran via MS Teams, were recorded and transcribed, and were analyzed through qualitative content analysis combining deductive and inductive coding. Feedback between cycles drove revisions, including added technical guidance on using the tools, and further modules were developed.

Five design principles derived from theory and feedback

The five principles emerged from the convergence of cognitive load theory, microlearning research, and the discussion findings. "Clear structure and orientation" calls for explicit learning objectives, an introductory overview, interlinked learning pathways, and summaries. "Guidance and navigation support" responds to technical friction through color coding, pictograms, and embedded usage instructions. "Segmentation and modularization" breaks content into small, self-contained units consistent with the transient information effect and with flexible, self-directed use. "Multimodal presentation" integrates video, text, and case examples, drawing on the modality and split-attention effects, since participants identified graphic structures such as decision trees as especially beneficial. "Learner activation" adds knowledge checks, reflection opportunities, and application tasks anchored in the worked example effect; participants also wanted feedback mechanisms and chances to apply content to their own teaching, which personalized feedback could only partially supply within the project's scope.

Platform choice and unresolved challenges

Technical design surfaced as the study's clearest practical problem. Participants found it hard to locate modules within OneNote, navigate backwards in Genially, and enable audio for embedded videos; one lecturer argued that no person can become well versed in five or twenty different tools. OneNote was widely judged too complex by lecturers and students alike and ill-suited as a learning-pathway environment, prompting suggestions of Padlet, Moodle, Taskcard, and Trello, with the flexibility of card-based platforms singled out as advantageous. Beyond platform selection, the author reports that choosing topics was difficult given the rapidly evolving nature of AI, which required continuous monitoring of content; that heterogeneous prior experience with digital tools produced elevated extraneous load for some participants; and that iterative refinement of the learning pathways proved resource-intensive. A proposed next step is aligning the modules with an established AI literacy framework.

What this means for practice

• Faculty developers. Treat the five principles as a checklist when building AI professional development: state objectives up front, group content into short self-contained units, and add visual navigation aids before adding media.
• Faculty developers. Pick the delivery platform deliberately — participants found OneNote too complex and favored card-based tools such as Trello, Padlet, and Taskcard — and provide explicit instructions for enabling audio and navigating embedded media.
• Instructors. Frame module content around core tasks (thesis supervision, competence-based assessment) and keep the emphasis on pedagogical, legal, and ethical dimensions rather than specific AI tools that quickly date.
• Instructors. Build in reflection and application tasks plus formative checks; the study's participants explicitly asked for feedback and opportunities to apply AI content to their own teaching.
• Institutions. Anticipate uneven digital competence among staff: differentiated technical support reduces the extraneous load that otherwise crowds out learning about AI competency.

Limitations

• The sample was very small: three lecturers in the first group discussion and four in the second, which restricts generalization.
• The researchers' close involvement with the field may have limited the critical distance maintained during the study.
• The group discussions frequently drifted into broad exchanges of subject-matter experience rather than a strict evaluation focus, which may have influenced the data.
• Not all features of microlearning could be implemented: smartphone optimization was out of scope and personalized feedback was incorporated only to a limited extent.

Citation

Cechovsky, N. (2026). Designing Microlearning Modules on AI for University Teachers in Vocational Teacher Education: A Design-Based Research Study. VETNET Conference Series.