Search Results (635)

Background/Objectives: AI chatbots may shape patient expectations before professional consultation. This scenario-based first-response study evaluated whether four user-facing chatbots provided orthodontic treatment-direction advice concordant with an expert benchmark and whether responses contained safety, referral, or overconfidence concerns. Methods: Forty fictional Turkish patient-oriented scenarios across eight categories were independently coded by three orthodontists as clear aligners, fixed appliances, both options, examination required, or advanced specialist/surgical evaluation required. Each scenario was submitted once to ChatGPT, Claude, Copilot, and Gemini on 20 May 2026. Two independent non-author orthodontists coded 160 archived first responses using a predefined framework, with adjudication before analysis. Results: Inter-expert agreement was moderate (Fleiss kappa = 0.491; Gwet AC1 = 0.528). Under the majority benchmark, exact concordance was 82.5% for ChatGPT, 67.5% for Claude, 42.5% for Copilot, and 37.5% for Gemini (Cochran Q = 34.105, p < 0.001). The overall difference remained significant in the 17 unanimous scenarios (Q = 11.455, p = 0.010), but a post hoc alternative-reference analysis that adopted the dissenting expert code in the 23 non-unanimous scenarios attenuated the rates to 57.5%, 52.5%, 52.5%, and 42.5%, respectively (Q = 4.222, p = 0.238). Coded safety-concern rates ranged from 15.0% to 62.5%. Conclusions: The sampled first responses differed in treatment direction and safety coding, but estimates were sensitive to the expert reference definition. Under the tested single-date, single-language, and single-run conditions, the findings represent a conditional snapshot rather than a time-invariant ranking of model capability. Patient-facing chatbots should support nondirective pre-consultation education and referral, not autonomous appliance selection. Full article

Motivation is a key predictor of success in second language (L2) acquisition, yet how culture and identity shape it among Latin American learners remains under-examined. This study used a quasi-experimental design with non-equivalent control and experimental groups (n = 100) to test the effect of an eight-week culturally responsive pedagogical intervention on the motivation of intermediate-level English-as-a-second-language learners at a public university in Ecuador, measured with the validated Spanish version of Gardner’s Attitude/Motivation Test Battery (AMTB). An independent-samples t-test revealed a statistically significant difference in post-intervention motivation scores between the experimental group (M = 156.1, SD = 31.2) and the control group (M = 134.5, SD = 28.4), t(98) = 3.91, p = 0.002, Cohen’s d = 0.72. The intervention was associated with a medium-to-large increase in motivation, consistent with sociocultural theory, the socio-educational model, the L2 Motivational Self System, and identity-investment theory; because the control group did not receive an equally novel activity, this finding should be read as preliminary evidence for the cultural/identity component specifically. Implications for culturally sensitive language teaching in diverse Hispanic contexts are derived. Full article

Background: Q&A corpora generated through university student–AI mental health support tools may reveal how physical activity (PA) and psychosocial themes are represented in support-oriented text. However, the absence of individual-level demographic metadata and the pooling of prompt and response fields limit attribution of any expression to a particular speaker, and the corpus describes a specific student population rather than a general or clinical one. Objective: This exploratory study described PA-, sport-, physical education (PE)-, body-, lifestyle-, and emotion-related patterns in a large corpus of university student–AI mental health exchanges collected through an institutional counselling platform. Methods: This study analysed 209,715 paired prompt–response records as combined exchange-level units using a BERTopic-based computational text-mining workflow. The full corpus was used for the main 18-topic model and overlapping dictionary analyses. After secondary data-quality filtering, 178,062 eligible exchanges formed the sampling frame from which a systematic sample of 10,000 exchanges was drawn for a separate complementary BERTopic and scenario-mapping analysis. The workflow used Qdrant/bge-small-zh-v1.5 embeddings, NFKC normalisation, an archived stop-word list, UMAP (n_neighbors = 15, n_components = 5, min_dist = 0.0, cosine metric, seed = 42), HDBSCAN (min_cluster_size = 300, min_samples = 10, Euclidean metric, EOM), c-TF-IDF topic representations, overlapping dictionary screens, and stability testing across seeds 42, 52, and 62. Results: A student/school/family-context lexical screen matched 83,215 exchanges (39.68%), and a broad PA/body/lifestyle screen matched 82,464 exchanges (39.32%). These overlapping indicators describe topical co-occurrence and do not establish PA behaviour or which party to the exchange produced a given term. Eighteen corpus-level themes were retained. In the 10,000-exchange analysis, 13.11% of exchanges matched a narrow movement-related expression screen, with the highest within-topic rate in the sample topic labelled emotional outburst and relaxation regulation (51.09%). Conclusions: The findings describe exchange-level lexical and topic patterns in student–AI interactions rather than actual PA behaviour, intervention delivery, clinical efficacy, or population prevalence, and they do not identify which party introduced the language. The mapping to autonomy, competence, relatedness, and emotional regulation is a post hoc interpretive lens, offered as a hypothesis to inform future, prospectively validated design work in PE and digital mental health support rather than as a demonstrated result. Full article

Generative AI (GenAI) adoption in higher education (HE) raises significant ethical concerns. The focus is shifting from rules- or outcomes-based learning environments towards the development of moral character, personality traits, integrity, and practical wisdom (phronesis). However, most existing AI ethics validation instruments are predominantly shaped by Global North paradigms. In Global South HE contexts, in particular, open distance e-learning (ODEL) HE institutions (HEIs) characterised by limited direct supervision and a digital divide, validation remains scant. Ethical risks are intensified by the adoption and integration of GenAI tools, such as large language models (LLMs), to enhance teaching, learning, research, and student support, thus recognising the need to develop and validate virtue ethics scales. The current paper attempts to address this gap by validating the Virtue Ethics Measurement Scale (VEMS) within South Africa’s largest comprehensive ODEL institution. Guided by the positivist paradigm, a 36-item cross-sectional survey of 503 undergraduate and postgraduate students measured six virtue dimensions (justice, honesty, responsibility, care, prudence, and fortitude). Confirmatory factor analysis (CFA) compared four competing models. The single-factor model showed poor fit, rejecting unidimensionality. A second-order hierarchical model demonstrated an acceptable fit (χ²/df = 2.992, CFI = 0.933, RMSEA (Root Mean Square Error of Approximation) = 0.063, SRMR (Standardized Root Mean Squared Residual) = 0.043) with subscale reliabilities ranging from Cronbach’s α = 0.84 to 0.90, supporting a multidimensional yet hierarchical virtue structure. The VEMS offers a psychometrically sound instrument for evaluating ethical AI use in ODEL institutions. This aligns with virtue ethics theory, which emphasises that moral character is a constellation of dispositions (e.g., honesty, care, prudence) rather than a single trait. The VEMS thus enables HEIs to assess students’ virtues, design targeted ethics capacity-development programmes, and inform policy reform for responsible GenAI adoption in under-researched Global South HE settings. Full article

Generative artificial intelligence (AI) and large language models (LLMs) are increasingly evaluated in rehabilitation, yet their clinical validity, reproducibility, and safety remain uncertain. This scoping review mapped peer-reviewed studies of generative AI/LLMs across rehabilitation assessment, clinical reasoning, decision support, planning, education, and functional classification. Following JBI methodology and PRISMA-ScR, five databases were searched for English-language studies published from 1 January 2015 to 24 July 2026. Two reviewers independently conducted study selection, data extraction, methodological appraisal, and application-domain coding. Of 2126 records, 43 publications representing 42 unique studies were included, predominantly from 2025–2026 and involving GPT/ChatGPT/OpenAI-family systems. At the unique-study level, six application domains were identified: clinical reasoning and decision support (n = 13), rehabilitation education, simulation, and feedback (n = 10), rehabilitation planning and prescription (n = 9), guideline adherence and clinical-question support (n = 6), adaptive feedback and rehabilitation support (n = 2), and assessment and functional classification (n = 2). Evidence was concentrated in benchmark, scenario-based, and educational evaluations, with limited patient-level outcomes. Heterogeneous methods, incomplete reporting of reproducibility, inconsistent safety assessment, and possible selective publication limited comparability and clinical generalizability. Generative AI/LLMs should therefore be used primarily as clinician-supervised assistive tools, with prospective validation, standardized reporting, and active safety evaluation prioritized. Full article

Over the past few years, a substantial amount of research has been conducted on the use of artificial intelligence (AI) in education. The latest AI achievements are very important for education because they expand access and can support personalization of learning at a new level. Personalized educational technologies leverage data analytics, machine learning, and adaptive learning algorithms to tailor educational content, learning pathways, and assessment methods to the individual needs, preferences, and abilities of learners. This review analyzes and classifies the recent advances in intelligent technologies that support the personalization of e-learning from a technological perspective. During our literature search, we used major academic search engines and digital libraries (such as Google Scholar, Scopus, IEEE Xplore, and Web of Science) together with bibliometric analysis tools to identify relevant and reputable scientific publications on the use of artificial intelligence in personalized e-learning. We comprehensively discuss the strengths, drawbacks, and the state of current use of each type of intelligent technology in personalized education. This paper also examines current trends in AI-driven personalized education, including the use of machine learning, learning analytics, adaptive assessment, ontologies, and Generative AI applications and tools. Furthermore, the study explores future directions in personalized learning, highlighting the potential of integrating advanced AI models and technologies to enable real-time personalized learning support. The findings suggest that AI-powered personalized educational technologies have the potential to significantly improve education by creating more inclusive, efficient, and learner-centered experiences. Full article

This study investigates the benefits of expanding small and open Large Language Models (Llama 3.x family) with retrieval and explicit referencing (retrieval-augmented generation), configured to refuse to respond when unsure about the answer. Additionally, we compare this augmented system with a cutting-edge commercial model (OpenAI GPT-4.5) in electrical engineering (EE) education questions. We build on an agentic RAG pipeline with retrieval from our team’s human-curated pedagogical reference document and a source-preserving, sectioned prompt that enforces citations. Six very experienced EE professors blindly assessed the answers to three exam-style questions of different types, generated by (i) the baseline model Llama 3.1 8B, (ii) our proposed RAG agent, and (iii) GPT-4.5, according to five dimensions: Correctness, Structure, Completeness, Precision, and Conciseness. The results show that the proposed RAG agent substantially improves the lightweight base model and produces transparent, syllabus-grounded answers that experts rated as correct and concise, while GPT-4.5 retains an edge on longer, multistep, and topology-intensive tasks. The local RAG agentic AI system achieved competitive performance relative to GPT-4.5 on dataset-scoped conceptual, procedural, and direct numerical electrical engineering circuit analysis educational tasks, which support the potential of curriculum-grounded open-weight models for constructing pedagogical applications. Full article

This study explores the educational experiences of Syrian Muslim refugee students in higher education in Türkiye, with particular emphasis on social integration, belonging, and identity construction in relation to religion and religious education. Using a qualitative research design, data were collected through semi-structured interviews with 19 participants and analyzed through content analysis, leading to the development of thematic categories. The findings indicate that these experiences constitute a non-linear, multi-layered, and dynamic process shaped by temporal and contextual influences. Language proficiency, social interaction, and the educational environment are identified as key determinants of social integration, which emerges as a fragile and reversible rather than linear process. The sense of belonging develops through individual experiences and does not necessarily align with social acceptance, pointing to a distinction between its internal and external dimensions. Identity construction is characterized as flexible, hybrid, and continuously negotiated. The study’s primary contribution lies in revising the initial conceptual model. While the mediating role of social integration is supported, religion and religious education are shown to function as context-dependent and ambivalent factors. Although religion may promote meaning-making, social connection, and belonging, it may also contribute to exclusion depending on the context. Its influence is largely indirect, operating through social integration. Overall, the study offers a more nuanced and dynamic model for interpreting refugee students’ educational experiences. Full article

Background/Objectives: As patients increasingly rely on large language models (LLMs) for Chronic Rhinosinusitis (CRS) diagnosis, surgical candidacy, and perioperative care, evaluating the accuracy of LLM-generated information against established clinical practice guidelines for surgical management of CRS is essential. Methods: ChatGPT, Google AI, Google Gemini, and Grok were queried using a 21-question guideline-mapped prompt set (long) and a single patient-focused prompt (short). Two physician reviewers independently scored responses using a 3-point rubric across 21 fields. Primary outcomes were guideline-concordant scores; secondary outcomes included readability measured with the Flesch Reading Ease (FRE) and Flesch-Kincaid Grade Level (FKGL). Inter-rater reliability (IRR) was assessed using the intraclass correlation coefficient (ICC). Analyses were performed in SPSSv31. Results: Guideline concordance ranged from 55.36% to 77.98% (p > 0.05), highest for Grok (77.98%, 95% CI 63.67–92.28), followed by Google Gemini (66.67%, 95% CI 35.42–97.91), ChatGPT (55.95%, 95% CI 29.25–82.65), and Google AI (55.36%, 95% CI 29.97–80.75), with Grok significantly outperforming both ChatGPT and Google AI. Prompt structure significantly affected scores. Long-form prompting resulted in higher guideline concordance scores than short-form prompting (+26.19, p < 0.001). The CPG demonstrated a more readable structure, with a higher FRE (44.1), exceeding scores generated by Grok (31.7), ChatGPT (39.9), Gemini (39.2), and Google AI (32.5). In contrast, the CPG was a higher reading grade level (FKGL score of 11.7) than Grok (11.4), Gemini (10.5), and ChatGPT (10.3), but was lower in reading grade compared to Google AI, which produced the highest FKGL score (12.5). IRR was high (ICC = 0.961). Conclusions: LLMs demonstrated similar guideline concordance, suggesting patients can expect comparable accuracy across platforms. While LLMs generally improved FKGL scores compared to the AAO-HNS CPG, they demonstrated lower FRE scores, indicating mixed results on overall readability. However, longer prompt structure meaningfully influenced output quality, highlighting how a user’s ability to frame precise prompts is critical to obtaining accurate information. Full article

Background: Large language models (LLMs) increasingly pass medical board examinations and aid clinical knowledge retrieval and decision support; validating their specialist knowledge is a prerequisite for safe use. Two limitations weaken existing evidence: many studies reuse public questions that may be in training data, and most evaluations use only English-language examinations. Plastic surgery serves here as a representative, highly specialized subfield of medicine. Methods: We used the non-public board examination of the Turkish Society of Plastic, Reconstructive, and Aesthetic Surgeons (TSPRAS): 100 single-best-answer (SBA) items in official parallel Turkish and English versions. Six contemporary LLMs from three developers, in matched free and paid tiers, completed six runs per language (7200 responses) under consistent English instruction. Accuracy, inter-run reliability, and cross-language and tier differences were analyzed. Results: All models exceeded the 60% passing threshold (71.5–86.4% aggregate), with almost perfect inter-run agreement (Cohen’s κ ≥ 0.843). Question language produced no significant difference after correction, and item difficulty correlated strongly across languages (Pearson r = 0.928); divergent items reflected content, not language. Paid variants offered only incremental advantages. Conclusions: Current LLMs showed no significant accuracy difference between the Turkish and English examination versions under a constant English prompt, supporting supervised study-aid use on this MCQ format. Full article

Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG)-based hybrid AI systems are increasingly transforming higher education, yet it remains unclear whether they are adopted through similar or distinct technology adoption pathways. This study compares the adoption of these two AI paradigms among postgraduate students using the Unified Theory of Acceptance and Use of Technology (UTAUT). It proposes the Adaptive Dual-AI Model (ADAM), which extends UTAUT by examining architecture-sensitive AI adoption within a unified framework. The model incorporates Technology Readiness (TR) and AI Awareness (AIA) as mediating variables and generational differences (Gen Z and Gen Y) as moderating factors. Data were collected from 639 postgraduate students enrolled in Saudi universities, of which 619 valid responses were retained after data screening. Partial Least Squares Structural Equation Modelling (PLS-SEM) was employed to evaluate the proposed model. The results indicate that the two AI paradigms follow distinct technology adoption pathways. Behavioural intention toward standalone LLMs was primarily associated with Performance Expectancy (PE) and Effort Expectancy (EE), whereas the adoption of RAG-based hybrid AI systems was more strongly associated with indirect relationships involving Technology Readiness and AI Awareness. Multi-group analysis further revealed that Gen Z learners exhibited stronger associations with LLM adoption, whereas Gen Y learners demonstrated stronger relationships involving RAG-based hybrid AI systems. These findings support the proposed architecture-sensitive perspective of ADAM, suggesting that differences in AI architecture are associated with distinct technology adoption patterns. From a practical perspective, the findings provide guidance for higher education institutions in selecting complementary AI technologies according to learning objectives, learner characteristics, and evidence requirements. More broadly, the study contributes to technology adoption research by extending UTAUT to heterogeneous AI ecosystems and offers practical insights for designing adaptive, personalised, and evidence-aware learning environments. Full article

Large Language Models (LLMs) show promise for supporting formative assessment by generating feedback on students’ written mathematical reasoning. However, practical use in educational settings remains constrained by the need to manually curate representative few-shot exemplars for prompt construction. This study examines whether unsupervised semantic clustering can automate few-shot exemplar selection for LLM-generated formative feedback in middle-school mathematics. As a controlled methodological feasibility study, we generated and refined 100 exam-realistic constructed responses for a Grade 7 inequality task aligned with middle-school mathematics standards. Student responses were embedded using Sentence-BERT, projected into a lower-dimensional space using Uniform Manifold Approximation and Projection (UMAP), and clustered with Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) to identify dominant reasoning patterns and ambiguous responses. Representative centroid and boundary exemplars from the resulting clusters were then used to construct few-shot prompts for the Llama 3.3 70B model, which generated feedback for the remaining 93 responses. Six independent mathematics instructors evaluated the AI-generated feedback using a structured 0–5 usability rubric. Across all instructor evaluations, 538 of the 558 instructor ratings (96.42%) were 3–5, representing feedback ranging from fair, requiring moderate edits, to excellent, ready to send. More specifically, 482 of the 558 instructor ratings (86.38%) were scores of 4–5, indicating feedback requiring no edits or only minor revisions. The remaining 20 ratings (3.58%) were scores of 0–2, while 13 unique feedback messages received at least one low rating. Across all instructor evaluations, 20 of 558 ratings (3.58%) were assigned scores of 0–2, while 13 unique feedback messages received at least one low rating. Qualitative analysis of low-scoring cases revealed recurring failure modes, including hallucinated completeness in concise solutions, failed arithmetic verification, and false logic flagging for atypical reasoning patterns. These findings suggest that clustering-based exemplar selection may reduce manual prompt-engineering effort while supporting usable LLM-generated formative feedback in a controlled mathematics setting. However, the present study does not compare clustering against alternative exemplar-selection strategies, and therefore conclusions should be interpreted as evidence of feasibility rather than comparative superiority. Full article

Background/Objectives: Artificial intelligence (AI), including large language models, chatbots, machine-learning systems, and hybrid tools, is increasingly used to support patient-facing eye-care communication. This systematic review evaluated the accuracy and effectiveness of AI-powered systems used for patient counseling, education, communication, referral/follow-up, and management support in optometry and related eye-care settings. Methods: A systematic review was conducted according to a predefined protocol and PRISMA 2020 reporting principles. Searches covered studies published from January 2020 to 31 May 2026 in PubMed, Embase, Web of Science, Cochrane Library, and IEEE Xplore. Eligible studies were original primary studies evaluating AI-supported tools for patient-facing counseling, education, question answering, treatment or medication guidance, triage, referral, follow-up, screening linked to management, or clinical decision support. Methodological quality was appraised using relevant JBI critical appraisal tools. Results: Thirty-nine studies were included in the qualitative synthesis. JBI appraisal indicated heterogeneous study designs and reporting quality, with common limitations related to simulated prompts or AI-output evaluations, variable comparators, and inconsistent outcome reporting. Evidence was dominated by patient-facing large language models and chatbot-based tools used for patient education, question answering, readability improvement, glaucoma and myopia counseling, diabetic retinopathy referral/follow-up, cataract education, oculoplastic and retinal-condition questions, and multilingual educational support. Study designs, AI models, prompting approaches, comparators, clinical topics, and outcome definitions varied widely, supporting narrative synthesis rather than quantitative pooling. Conclusions: AI-powered systems show potential as supervised adjunctive tools for eye-care counseling, education, and management-related communication. However, evidence remains heterogeneous and dependent on simulated prompts, AI-generated outputs, and model-based evaluations. Future research should prioritize standardized evaluation, real-patient validation, safety monitoring, readability control, and patient-centered outcomes before routine implementation in optometry and broader eye-care practice across diverse clinical settings. Full article

Digital Artificial Intelligence (AI), exemplified by Large Language Models (LLMs) such as ChatGPT, has achieved remarkable progress across a wide range of applications, driven not only by advances in algorithms but also by the emergence of a shared research ecosystem built upon commodity computing platforms, standardized software frameworks, open-source models, benchmark datasets, cloud infrastructure, and broadly accessible educational resources. In contrast, Autonomous Vehicles (AV), AI systems that perceive, reason, and act in the physical world, have advanced more slowly despite substantial public and private investment. Progress remains constrained by fragmented research and educational infrastructure that limits reproducibility, interoperability, scalable validation, and workforce development. This paper surveys the current state of the AV research ecosystem, including hardware platforms, autonomy software stacks, datasets, simulation environments, digital twins, testing and validation frameworks, and educational programs. Drawing lessons from the evolution of Digital AI, the paper identifies key gaps in accessibility, standardization, integration, and openness across the AV technology stack and outlines opportunities to develop shared research testbeds, modular open platforms, interoperable software and data ecosystems, common benchmarks, and interdisciplinary educational programs that can accelerate autonomous vehicle innovation. Finally, the paper provides a framework for evaluating AV research and educational infrastructure which identifies priorities for future investment. Full article

Professional online communities increasingly function as digital social infrastructures during societal disruption, yet less is known about how their support practices reorganize during prolonged crises. This study examined how social support, engagement patterns, and support-seeking behavior changed in a large Hebrew-language professional Facebook group for Israeli educators across routine and crisis periods. Quantitative content analysis was conducted on an analytic sample of 1008 posts published between January 2022 and July 2025, covering the pre-war period and the wartime period following October 2023. Posts were coded for informational, emotional, and instrumental support and for support-seeking, and negative binomial and logistic regression models examined engagement and predictors of support-seeking. Informational support was dominant (69.3%), followed by support-seeking (30.7%), emotional support (12.8%), and instrumental support (4.4%). During wartime, support-seeking and instrumental support were less frequent, whereas informational and emotional support were more frequent. Informational posts received approximately three times more likes, while support-seeking posts received fewer likes but sustained comment activity. Anonymous authors were more likely to seek support, but posts by anonymous authors received fewer comments. The findings suggest an aggregate pattern of support flexibility in the analytic sample, reflected in a relative shift from reactive support-seeking toward proactive informational sharing and emotional processing as part of collective resilience-building during sustained societal crisis. Full article