A novel gross deletion in the progranulin gene in four subjects with frontotemporal dementia.

Large language models (LLMs) have rapidly permeated higher education, offering new possibilities for academic writing, exam preparation, and personalised learning []. In dental education, where evidence-based practice and critical thinking are paramount, the integration of LLMs presents both opportunities and challenges []. Dental students are increasingly adopting LLMs for academic tasks, but their usage patterns, perceived academic outcomes, verification behaviours, and risk perceptions appear to be shaped by a complex interaction of ethical concerns and institutional guidance []. Early scholarly and policy-oriented discussion around LLMs has understandably centred on assessment integrity, academic misconduct and potential risks to critical thinking and independent judgement [,,]. However, multiple reviews and surveys indicate that empirical evidence is still developing, uneven across contexts and often dominated by cross-sectional, single-setting or single-country samples, which limits what can be inferred about generalisable versus context-specific patterns [,]. In parallel, analyses of institutional AI documents suggest that institutional responses have frequently prioritised academic honesty language over more comprehensive implementation and review mechanisms, contributing to variability in what students are told and how use is governed []. Moreover, educational consequences of LLM use are unlikely to be binary. Students can report strong perceived usefulness and efficiency while simultaneously acknowledging ethical concerns and reliability risks, particularly in assessment-related contexts []. Evidence from medical and dental education suggests that responsible use requires more than measuring frequency of use [,,]. It also requires attention to students’ purposes for using LLMs, perceived benefits, and verification routines for managing hallucinations, fabricated citations, and plausible-sounding misinformation. This also implies that understanding student practice requires examining how LLM outputs are evaluated against conventional academic resources and credible evidence, rather than treating LLMs as standalone answer tools [,].