instructional design

This study addresses the misalignment between traditional software engineering education—still reliant on manual coding assessments—and contemporary industry practices shaped by the widespread adoption of large language models (LLMs). To bridge this gap, the authors propose a theoretical framework for LLM-era software engineering education that shifts the focus from code production to critical validation, human-AI collaboration governance, and process transparency. The framework introduces a novel pedagogical model integrating LLMs into curricula, specifically designed for high-control, large-enrollment, exam-oriented educational contexts such as computer engineering programs in Turkey. It redefines academic integrity by prioritizing transparent development processes over conventional plagiarism detection and offers a foundational theory for curriculum reform, while calling for longitudinal empirical studies to validate its efficacy.

This work proposes a novel approach that integrates large language models with interactive concept visualization to support K–12 teachers in efficiently generating structured, editable scaffolding questions aligned with engineering design processes. Recognizing that educators often spend considerable time manually crafting guiding questions to help students navigate complex design challenges, the method automatically parses task requirements, decomposes core engineering concepts, and produces pedagogically appropriate inquiry prompts. By synergistically combining the generative capabilities of large language models with visual, manipulable concept representations, this framework enables teachers to rapidly construct accurate, contextually relevant instructional supports tailored to each phase of the engineering design cycle, thereby significantly enhancing lesson preparation efficiency without compromising content fidelity or pedagogical suitability.

In LLM-augmented education, students often exhibit passive learning behaviors and overreliance on large language models, hindering deep conceptual understanding and metacognitive development. Method: This paper proposes the “Student-as-Teacher” paradigm, wherein students actively design programming problems embedded with *knowledge breakpoints*—intentional gaps in prerequisite knowledge—and guide LLMs to solve them iteratively. We implement Socrates, a lightweight interactive system supporting breakpoint annotation, progressive prompting, and feedback-driven refinement loops. Contribution/Results: By reversing the LLM’s role—from tutor to tutee—the paradigm transforms problem authoring into a high-order learning activity that fosters active engagement and self-regulated learning. A controlled study in undergraduate computer science courses demonstrated statistically significant improvements: +12.3% in final exam scores and enhanced conceptual mastery, validating its efficacy in promoting active learning, reducing tool dependency, and strengthening metacognitive awareness.

This study addresses a critical tension in engineering education: students’ simultaneous expectation of efficiency from large language models (LLMs) and their frequent misjudgment of LLM capabilities, which can lead to accuracy errors, bias propagation, academic integrity risks, and overreliance. Through a mixed-methods approach combining surveys, qualitative metaphor analysis, and critical literature review, the research reveals that students primarily employ LLMs for writing assistance, conceptual clarification, programming support, and brainstorming. It identifies a pervasive tendency to overattribute authority to LLMs and underappreciate the burden of verification. Introducing the metaphors of “oracle” and “mentor” alongside the concept of “cruel optimism,” the study elucidates the gap between students’ role expectations and LLMs’ actual capacities. It proposes an integrative framework centered on goal-directed use, contextual sensitivity, critical AI literacy, reflective evaluation, and ethical scaffolding to guide the responsible integration of LLMs in engineering education.

Students lack structured, critical, and practice-oriented large language model (LLM) literacy. Method: We redesigned a post-CS1 course to systematically integrate four modules—LLM fundamentals, hands-on tool usage, ethics discussion, and reflective practice—and introduced a dual-strategy framework: “disclosure mechanisms” (requiring transparent AI use) and “validation training” (cultivating systematic verification of LLM outputs). Pedagogical approaches included LLM toolchain exercises, interactive classroom activities, pre-/post-intervention surveys, explicit critical thinking instruction, and an AI-augmented problem-solving framework. Contribution/Results: Students demonstrated deeper technical understanding, more deliberate and ethically grounded LLM usage, significantly enhanced validation awareness, and improved human-AI collaboration competence. The curriculum design is discipline-agnostic and offers a scalable, responsible pedagogical paradigm for computing education in an AI-integrated era.

Current LLM-assisted programming education systems lack a unified conceptualization of their assistance boundaries, implementation approaches, and control mechanisms, hindering education-oriented comparative analysis. Addressing this gap, this study conducts a scoping review and qualitative content analysis of 90 relevant systems, applying thematic coding to develop an innovative three-dimensional governance framework—PEA (Policy, Enforcement, Authority)—alongside a corresponding governance codebook. The framework reveals pervasive issues in existing systems, notably excessive centralization of authority and insufficient configurability. This work provides the first systematic mapping of governance design patterns in LLM-based programming support tools, establishing a theoretical foundation and a structured design vocabulary for developing next-generation educational tools that are goal-aligned, configurable, and accountable.

This work addresses the growing disconnect between software engineering education and industry practices, as academic curricula have lagged behind the rapid adoption of emerging technologies such as large language models (LLMs) and Model Context Protocol (MCP). To bridge this gap, the study introduces a novel pedagogical framework that synergistically integrates LLMs and MCP into the curriculum through intelligent programming assistance, engineering simulation platforms, and university–industry collaborative internships. This approach fosters a collaborative learning environment closely mirroring real-world industrial workflows. Empirical results demonstrate significant improvements in students’ programming proficiency, complex problem-solving capabilities, and competence in leveraging AI-powered tools. By aligning academic training with contemporary industry demands, this initiative effectively narrows the theory–practice divide and catalyzes a paradigm shift in software engineering education for the AI era.

This study investigates how to effectively support learners in formulating high-quality questions across computational education tasks with varying degrees of openness. To this end, two large language model–based scaffolding systems were designed and pilot-deployed: guided questioning (an indirect scaffold) and worked examples (a direct scaffold), evaluated through a unified framework grounded in Bloom’s taxonomy. The work innovatively compares and integrates these two AI scaffolding modalities, proposing a sequential “reflect-then-exemplify” strategy that balances immediate question quality improvement with deeper cognitive engagement. Findings indicate that direct scaffolding yields more pronounced immediate gains in question quality, whereas indirect scaffolding is more favorably received by students and fosters richer reflection on the process of question design.

This study addresses the low completion rates and poor learning outcomes among students with low prior knowledge or motivation in traditional programming exercises, which often lack personalization. To tackle this issue, the authors propose a learner-profile-driven approach that integrates large language models with the FACET educational framework for the first time. Guided by Bloom’s taxonomy of cognitive processes and Self-Determination Theory, the method dynamically adapts scaffolding structures, instructional explicitness, and linguistic tone to enhance task alignment while preserving desirable difficulty. Experimental results demonstrate that the proposed approach achieves over 99% task completion across all students—compared to 25–30% non-completion among low-knowledge students in the control group—and improves accuracy by 18.2% for learners with low knowledge or motivation, without significantly increasing perceived challenge.

This study investigates how intelligent tutoring strategies based on large language models (LLMs) can enhance students’ ability to independently and effectively use LLMs in programming education. Comparing Socratic guidance (SG) with prompt refinement (PR), the research employs educational experimentation and behavioral analysis to provide the first empirical evidence that, although SG is perceived as less efficient, it more effectively fosters understanding-driven prompting strategies among learners. In subsequent unconstrained interactions with LLMs, students in the SG condition demonstrated significantly higher learning gains and superior prompting behaviors compared to those in the PR condition. These findings highlight the critical role of guided dialogue in cultivating long-term autonomous learning capabilities with AI tools.