Adaptation of Social Health Care Tools From Paper-Based to Web-Based Formats: Scoping Review.

Large language models (LLMs), such as OpenAI’s ChatGPT series [], have rapidly transformed how information is accessed, synthesized, and applied across educational and professional domains [-]. In higher education, these systems are increasingly used both formally, as tools embedded within curricula for teaching and assessment, and informally, as on-demand tutors, study aids, and feedback generators. In the health sciences, including dentistry, LLMs may help learners master complex theoretical concepts, refine procedural skills, and prepare for clinical practice. They may also assist educators with lecture development, assignment scoring, and test construction []. However, the quality, tone, factual orientation, and presentation of LLM-generated responses can vary substantially depending on how prompts are formulated []. This variability highlights the importance of prompt engineering, defined as the deliberate design of instructions to guide AI systems toward desired outputs []. Research on prompt engineering shows that even small changes in a prompt’s wording, structure, or perspective can alter the characteristics of the resulting output [,]. Previous studies have demonstrated that carefully designed prompts can improve the clarity, relevance, and depth of AI-generated responses [,], and in some contexts, enable AI-generated feedback to outperform feedback from novice humans []. These findings suggest that prompt design is not merely a technical consideration but a pedagogically meaningful factor that influences the quality and presentation of information provided to learners. Despite these advances, most existing work has examined prompt engineering in broad academic contexts [], with limited attention to highly specialized professional education, such as dentistry, where accuracy, safety, applicability, and evidence-informed communication are particularly important. Prompt framing refers to the perspectives or priorities embedded in prompt wording []. Framing effects are well established in behavioral science, in which different emphases can alter attitudes and decisions even when the underlying factual content is similar []. In LLMs, framing may shape response content, tone, stance, and elaboration []. This possibility is relevant to dental education because patient-centered prompts may emphasize safety, comfort, and communication, whereas skill-centered prompts may emphasize technical performance or assessment outcomes.