The rapid integration of Generative AI (GenAI) into higher education presents both opportunities and systemic challenges, particularly in domains where feedback is central to learning. This study investigates the capacity of a large language model to generate formative feedback for student-created UML diagrams in a university software engineering course. Across two cohorts (N = 262), AI-generated, teacher-generated, and no-feedback conditions were compared, analyzing student perceptions, learning outcomes, and grading reliability. Results show that while students rated GenAI feedback as beneficial and often comparable to human comments, teacher feedback remained more effective in supporting performance gains, especially in complex modeling tasks. Linguistic analysis further revealed that GenAI feedback was more repetitive and less pedagogically rich than human feedback. Beyond these course-level findings, the study highlights broader implications for higher education systems. GenAI feedback represents not just a pedagogical tool but a cognitive partner that can reshape assessment models, curriculum design, and faculty roles. Its scalable nature offers potential to democratize access to high-quality formative feedback, while also raising equity, accountability, and policy challenges at institutional and sectoral levels. By situating empirical results into this broader frame, the study argues that GenAI is catalyzing a paradigm shift toward new systems of learning, where feedback becomes systemic, scalable, and embedded in the very structure of higher education.

Exploring pathway scenarios for future urban water resilience using systems approach coupled with an agent-based model operated on synthetic social-policy networks.

Effective verbal feedback is crucial for the development of medical trainees; however, the lack of a standardised feedback evaluation framework makes the delivery of effective feedback challenging for faculty. To develop and refine a feedback evaluation rubric based on the Self-Assessment, Feedback, Encouragement, Direction (SFED) model that can be used as a versatile tool to guide faculty in providing and evaluating verbal feedback. We created an evaluation rubric to assess faculty feedback and provide a model for evaluating verbal feedback to a trainee based on the SFED model, which promotes formative, actionable, and learner-engaged feedback. The rubric was designed by clinicians and a medical education specialist who was an expert user of feedback models. It features the four SFED categories, each delineated into 2-3 subcategories scored on a 1-5 ordinal scale, supplemented by a global judgement score. Rubric refinement involved applying the rubric to anonymous verbal feedback transcripts to improve scoring delineation across the scale. Preliminary validation of the rubric was performed with three independent, trained raters evaluating transcripts from verbal feedback sessions and consistency was evaluated using an Intraclass Correlation Coefficient (ICC). This versatile tool can be utilised for real-time observation, transcript analysis, or as a proactive planning tool to train faculty to provide feedback. Preliminary validation of the rubric demonstrated an excellent inter-rater reliability with an ICC of 0.90. The application of the SFED rubric has the potential to improve the quality of verbal feedback delivered to trainees, fostering trainee growth, and ultimately enhancing clinician development.

Designing conversational Agents for adaptive instructional support in business simulation gaming

Can students judge like experts? A large-scale study on the pedagogical quality of AI and human personalized formative feedback

Evaluating adaptive and generative AI-based feedback and recommendations in a knowledge-graph-integrated programming learning system

cervical nuclei segmentation through synergic conditional generative adversarial network in cervical smear images.

Neurologic disorders in the neonatal period are a leading cause of death and disability in low- and middle-income countries, including Latin America. Access to specialized neonatal neurocritical care remains limited, and educational opportunities are scarce. We aimed to evaluate the feasibility, reach, knowledge acquisition, and perceived value of a virtual educational series in neonatal neurology targeting health care professionals across Latin America. This prospective educational intervention included 6 case-based webinars delivered between October 2023 and April 2024 in Spanish via Zoom. The series, organized by the Newborn Brain Society's Latin American Task Force, was open-access and promoted through regional medical societies. Participants completed demographic surveys, pre- and postsession knowledge assessments, and standardized feedback forms. The virtual webinar series covered key neonatal neurologic topics, including hypoxic-ischemic encephalopathy, seizures, hypotonia, and neuromonitoring. Presenters were local trainees with expert moderator commentaries. Outcomes included webinar attendance, knowledge acquisition (multiple-choice questions), and satisfaction (5-point Likert scale). Responses were analyzed overall and for matched pre- and postsession participants. A total of 1,424 participants from 24 countries attended the 6 webinars, with each session exceeding 100 attendees and 10 countries (meeting predefined feasibility criteria). Attendees were predominantly physicians (84%), although nurses, residents, and other professionals also participated; 23% attended multiple sessions. Knowledge scores improved significantly, with an 8%-18% increase in matched pre- vs postsession scores (p < 0.05). Participant satisfaction was high (mean feedback ∼4.8/5), and qualitative feedback highlighted the series' clinical relevance, interactivity, and value for practice. This virtual educational initiative proved feasible, engaging, and effective in improving neonatal neurology knowledge across Latin America. Such tele-education programs may help reduce regional disparities in care.

ABSTRACT First-year financial accounting students face high content volume and limited access to timely formative feedback, a gap that scalable digital tools could address. This study examines whether voluntary engagement with a fully gamified self-testing app ‘DUGA’ improves exam performance, and which dimensions of engagement matter most. Data were collected across two independent cohorts (N = 1,248) at an open-access European university using a quasi-experimental design. We compare exam performance between voluntary app users, rigorously controlling for prior accounting ability. We decompose engagement into two distinct indicators: intensity (quantity of questions attempted) and accuracy (proportion answered correctly). Results show that voluntary app users significantly outperform non-users. Among users, accuracy is strongly, positively associated with exam performance, whereas intensity alone is not. Accuracy fully mediates the intensity–performance relationship, indicating that it is not engagement volume itself, but the quality of engagement it produces, that drives exam performance. Moreover, intensity and accuracy combine synergistically: students who practise both frequently and accurately achieve greater exam gains than accuracy would predict independently. These findings help explain mixed evidence in prior gamification research by distinguishing engagement quality from engagement quantity and offer practical guidance for accounting educators on the design and implementation of gamified self-testing tools.

Youth exposed to household challenges are among the groups most vulnerable to early substance use. To address the needs of this group, this pilot study assessed the feasibility of participant recruitment and retention, as well as the acceptability of a novel substance use prevention intervention, Better Together, for Black youth exposed to household challenges. Participants were recruited using school and community presentations, digital flyers, and referrals. One-hundred twenty-seven students from two schools were screened. Participant enrollment, session attendance, and assessment completion were used to determine feasibility. Quantitative and qualitative data were collected after each session; responses were used to assess intervention acceptability. Demographic and substance use data were collected at baseline, post-test, and 1month follow-up. Fifty-nine youth were eligible, returned their parental permission forms, enrolled, and randomized. Nearly 90% of participants (n = 47) attended at least 5 sessions; the average number of sessions attended across conditions was 6.5. Fifty-four youth (92%) completed 1-month follow-up assessments. Session feedback forms showed that most students were happy after each session (61-85%). Qualitative feedback suggests that students were engaged in discussing the key messages using their own words. There were no significant changes in substance use knowledge or intentions over time or by condition. However, there was a small yet significant increase in substance use behaviors over time. Better Together may be a promising intervention to prevent substance use among Black youth exposed to household challenges. Findings suggest that recruiting, retaining, and engaging participants in the eight-session intervention is possible. Findings will inform plans to implement a full-scale study to evaluate the efficacy and contextual factors affecting program implementation. Clinicaltrials.gov: NCT06594835. Registered: September 10, 2024 - Retrospectively registered, https://clinicaltrials.gov/study/NCT06594835?locStr=Baltimore,%20MD&country=US&state=Maryland&city=Baltimore&intr=Better%20Together&viewType=Card&rank=1.

Medical education relies on narrative feedback to improve and assess medical student performance in clinical settings, but ensuring sufficient quality narrative feedback remains challenging. This study examined the impact on quantity and quality of narrative feedback of a "two-step process" in which students first documented verbal feedback received from faculty, and then those same faculty reviewed and modified the feedback recorded by the student. The two-step process was implemented at Baylor College of Medicine (BCM) in the context of planned direct observations of clerkship students and at Harvard Medical School (HMS) in the context of supervised clinical care by advanced elective students. Narrative feedback after the initiation of the two-step process (2023-2024) was compared to the baseline period (2022-2023). Differences in quantity of narrative feedback were measured by comparing word count and proportion of completed feedback forms (for supervised clinical care at HMS only). Quality of narrative feedback was measured by coding for and comparing markers of quality: behavior-based feedback, specific clinical examples, and absence of trait-based language. Using the two-step process was associated with increased narrative feedback at both institutions. Median word count per instance of feedback increased from 23 to 50 (P < .001) at BCM and from 75 to 119 (P < .01) at HMS. At HMS, the proportion of completed feedback forms also increased from 42 to 93% (P < .001). The proportion of narrative feedback instances with behavior-based feedback and specific examples significantly increased at both institutions, and use of trait-based language significantly decreased at HMS (P < .05 for all). A two-step process in which students and faculty co-create narrative feedback is feasible and enhances quantity and quality of feedback in various clinical settings. Future research should explore how student participation in generating narrative feedback impacts participating students and faculty.

This study developed and validated Echoes of Iriga: A Localized Literature-Based Supplementary Material in English 7 for Grade 7 learners of San Juan Integrated School, Iriga City, School Year 2025–2026. It addressed the need for culturally relevant, context-based, and learner-centered instructional resources that connect English learning with Irigueño heritage, community experiences, and local identity. The study was anchored on Culturally Responsive Teaching Theory, Schema Theory, Language and Culture Interdependence Theory, the researcher’s Peripheral Participation and Engagement Theory, and the Input–Process–Output model. Using a Research and Development design guided by the ADDIE Model, the study involved 51 Grade 7 learners selected through complete enumeration and 9 subject experts selected through purposive sampling. Data were gathered through a researcher-made questionnaire, an LRMDS validation checklist, and qualitative feedback forms, and were analyzed using mean, grand mean, and Likert scale interpretation. The developed material was based on three localized sources: Iriga: The Travails of a City’s Historic Odyssey, IRIGA Essays, and the Iriga City Official Website. Findings showed that learners rated the material highly in engagement, relevance, clarity, and aesthetic value, with aesthetic value receiving the highest grand mean of 4.92. Expert validators rated the material very satisfactory, with an overall grand mean of 3.95. The study concluded that the material is valid, culturally meaningful, well-designed, and appropriate for English 7 instruction. It supports SDG 4 – Quality Education and SDG 11 – Sustainable Cities and Communities by promoting contextualized learning and preserving local culture. Its sustainability impact lies in strengthening educational, cultural, community, and institutional sustainability through localized instructional material development.

Current robotic prostheses developed for individuals with transradial amputation often lack physiological feedback, particularly proprioceptive information, which limits control precision and increases reliance on vision. This study investigates the effect of introducing non-invasive somatotopic feedback using transcutaneous electrical nerve stimulation (TENS) to convey hand aperture information, as an artificial form of proprioceptive feedback. Twenty healthy participants with intact limbs were divided into two groups: one receiving TENS-based feedback and one without feedback. Participants performed an aperture control task under visual and non-visual conditions, with some trials including a concurrent Stroop task to assess cognitive load. We hypothesized that providing non-invasive somatotopic proprioceptive feedback via TENS mitigates key control and integration challenges, leading to improved accuracy, faster learning, and reduced reliance on vision, without increasing cognitive demands. Under visual deprivation, participants receiving TENS feedback achieved significantly smaller aperture control errors than those without feedback, both under alternating visual conditions (p < 0.001) and under prolonged visual deprivation conditions (p = 0.017). From the very first trial, TENS-based feedback enabled control accuracy comparable to that with visual input. Although perceptual shifts affected control accuracy under dual-task conditions, the +TENS group maintained high cognitive performance and effective control toward perceived functional targets, suggesting that the artificial feedback nonetheless supported hand control in the presence of a secondary task. These findings highlight the potential of noninvasive somatotopic proprioceptive feedback delivered through TENS to provide physiologically meaningful information and support future strategies for restoring proprioceptive function in prosthetic hand users.

School-based multicomponent interventions have shown promise in improving children's lifestyle behaviours, but evidence on their feasibility in Asia remains limited. The Promoting hEalthy Diet and Active Lifestyle intervention was developed for primary school children (ages 10-11 years) in Singapore to improve diet and activity behaviours (fruit and vegetable intakes, moderate-to-vigorous physical activity (MVPA) time, and recreational screen time), through lessons, take-home activities, parental engagement, and environmental modifications. This pilot study evaluated six feasibility dimensions: recruitment; data collection; implementation fidelity; practicality; social validity; and potential effectiveness, particularly of our theory-based approach, which focuses on ideations and their impact on behaviour. A concurrent mixed-methods study design was employed in two public co-educational schools. This involved a pre-post survey of students' self-reported behaviours (n = 139) and to generate healthy eating (HE) and physically active (PA) ideation scores, focus group discussions with students (n = 20) and teachers (n = 11), teacher-reported feedback forms (n = 13 classes), short interviews with canteen vendors (n = 11), and observation to evaluate environment modifications. Student survey completion rate was low (57%) due to insufficient time, although lessons were well-liked by students and teachers, with high student attendance (97%). Teachers delivered 83% of the planned activities, with time constraints being a barrier. There were no significant changes in the behaviours or HE ideation scores, although PA ideation scores improved (+ 0.3, 95% CI: 0.03-0.6) due to increased knowledge scores (+ 0.5, 95% CI: 0.4-0.7). While this increase in knowledge did not result in behavioural changes, the association between ideation scores and behaviours aligned with theoretical expectations. Higher HE ideation scores correlated with greater fruit and vegetable intake, while higher PA ideation scores were associated with more MVPA and less recreational screen time. Poor fidelity in take-home activities and parental engagement, along with students' inertia and perceived lack of need for change, explained the absence of significant behavioural change. The study identified key conditions for better implementation, including stronger stakeholder buy-in, flexible lesson delivery, and digital data tracking. It highlighted the complexity of behaviour change in Singapore primary schools, emphasising the need for long-term, tailored interventions. ISRCTN16114046. Registered 16/10/2022, prospectively.

ABSTRACT This paper introduces GenAI-by-Design, a theoretically grounded, research-informed pedagogical framework for integrating Generative AI (GenAI) into higher education instruction and assessment. Grounded in the pedagogy Learning by Design and UNESCO’s AI Competency Framework for Students, the approach is also informed by results from mixed-methods studies conducted in graduate-level education programmes in the USA. Specifically, these studies explored the affordances and limitations of GenAI in providing formative feedback alongside human commentary. Initial findings revealed students’ overwhelming preference for human feedback. Nevertheless, after implementing a calibrated AI system, perceptions shifted markedly. Student and instructor participants characterised GenAI as a valuable feedback source, welcoming its addition to formative assessment. Findings also pointed to development in students’ AI literacy and critical engagement with GenAI tools. The framework operationalises these insights through scaffolded knowledge processes and competency progression, emphasising ethical, dialogic, inclusive, student-centred learning while safeguarding equity, transparency, and human agency in AI-supported learning ecosystems.

ABSTRACT Grounded in interactive models of second language (L2) listening, schema theory, sociocultural accounts of mediation, and control–value frameworks of achievement emotions, this study examines a theory-informed AI-supported approach to listening instruction. In a Saudi university English as a foreign-language (EFL) programme, 66 first-year learners in two intact sections participated in a quasi-experiment. Both groups followed the same syllabus, audio texts, and instructor; however, the experimental group used adaptive pre-listening prompts, learner-controlled pause and replay (with moderated pacing), strategic on-demand transcripts, and brief post-listening dialogue with immediate formative feedback. Listening comprehension was assessed with parallel tests, listening anxiety with the Foreign Language Listening Anxiety Scale (FLLAS), and flow with an adapted classroom flow scale. Mixed-design ANOVAs showed significantly greater comprehension gains for the AI-supported group (post-test d = 0.88; gain differential d = 1.30), with particularly strong effects for inferencing and contextual lexical inference. Listening anxiety decreased substantially in the AI-supported group, whereas the comparison group showed minimal change, and post-intervention flow was higher for AI-supported learners. Anxiety reduction was moderately associated with listening gains (r ≈ −0.42). Overall, the results suggest that a carefully designed AI-supported listening cycle can improve listening development while also reducing anxiety and supporting sustained engagement.

• We examine travellers’ choices between autonomous vehicles (AVs) and human-driven vehicles (HVs) under different information conditions. • A repeated mode-choice game was implemented where participants made decisions based on varying levels of information about travel outcomes. • Mixed multinomial logit (MMNL) models were estimated to analyse the factors influencing travellers’ choices. • Learning patterns, heterogeneous preferences, and fairness concerns significantly influence mode choice in mixed AV-HV traffic environments. With rapid advancements in automation, autonomous vehicles (AVs) will soon coexist with human-driven vehicles (HVs) in mixed traffic. This study investigates the coordination behaviour in mode choice between AVs and HVs with positive and negative externalities, respectively. Participants play a mode choice game repeatedly under partial information (PI) or full information (FI) feedback on payoffs of the alternatives. The descriptive analysis indicates that choices under PI are close to the outcome at the Nash equilibrium; however, choices under FI are consistently below this equilibrium outcome. This discrepancy leads to a decline in social welfare, despite the provision of additional information. To test the factors influencing choice behaviour, we built a series of mixed multinomial logit (MMNL) models. The first model reveals substantial heterogeneity in individual preferences for AVs and indicates that full information feedback weakens the influence of inherent preferences while reinforcing the role of dynamic experience learning. In the second model, the memory effect diminishes further when payoff fairness is introduced. Fairness considerations significantly shape mode choices: while some participants prioritise equitable payoffs, others pursue individual payoff maximisation, producing aggregate outcomes between the distributions induced by the payoff equity and by the Nash equilibrium. By understanding the factors influencing choice behaviour and payoff preferences, policymakers and transport planners can develop targeted interventions to promote AV adoption, ultimately facilitating a smoother transition towards autonomous vehicle usage.

We investigate the role of baryonic physics in shaping the population, structure, and internal dynamics of galactic subhalos using the Mochima suite of cosmological zoom-in simulations. A refined method is developed to identify bound subhalo material by isolating the local gravitational potential and applying multi-criteria phase-space selection. This approach enables a robust characterisation of subhalo properties across five baryonic runs with varying prescriptions for star formation, and supernova and protostellar feedback, as well as a dark matter-only baseline. At the population level, we find that the concentration of the central massive host halo, modulated by baryonic physics in the central disc galaxy, is a key predictor of subhalo survival. Subhalos with more massive stellar components exhibit deeper internal potentials and enhanced resilience to tidal disruption. At the structural level, we identify a broad diversity in inner dark matter profiles, consistent with observations of dwarf galaxies. We show that this diversity correlates with both star formation history and environmental interaction. In particular, galaxies that form most of their stars early tend to retain steep cusps, while those with extended or recent star formation exhibit oscillating inner slopes shaped by bursty feedback and tidal perturbations. These findings suggest that the so-called “diversity problem” may reflect the complex interplay between feedback history and gravitational environment, rather than a breakdown of cold dark matter predictions.

Paramedic involvement in general practice has expanded rapidly in response to workforce pressures and rising patient demand. Feedback and educational support are recognised as essential for clinician development, wellbeing, and patient safety. While existing studies have explored paramedics' transition more broadly, little is known about the experiences of non-roadmap paramedics (NRPs), who fall outside structured supervision frameworks. To explore how NRPs experience feedback and educational support in primary care, and how these processes influence professional learning, role integration, and wellbeing. A qualitative study was conducted with eight NRPs working in English general practice. Participants were recruited via purposeful and snowball sampling. Semi-structured interviews were undertaken online, transcribed, and analysed thematically using Braun and Clarke's framework. Reflexivity and supervision supported rigour and credibility throughout. Four themes were identified. First, participants described adjusting from ambulance service feedback, based on negative events, to a more educational model in general practice. Second, ad hoc case-based discussions were the most common form of feedback, valued for their immediacy and reassurance. Third, time and organisational constraints limited structured learning opportunities, with formal appraisals often absent or superficial. Finally, autonomy was both empowering and isolating, influencing opportunities for feedback and role clarity. Supportive practice cultures fostered feedback as a routine and developmental process, aiding integration. Feedback is central to NRPs integration and professional growth in general practice. However, reliance on ad hoc exchanges and lack of structured support risk undermining development. Sustainable integration requires time, mentoring, and educational investment to normalise feedback as a shared responsibility.

Integrating Medical Writing into Pharmacy Curricula: Design and Evaluation of an Online Elective Course.