Oct 2026· Journal of Civil Engineering Education· Vol 152· 0 citations· 14 references
Abstract
Large language models (LLMs) are rapidly entering civil engineering research and practice, yet little is known about their use in educational contexts. This study reports results from an institutional case study based on a taxonomy-aligned survey of 109 respondents (103 undergraduates, four graduate students, and two faculty) in civil engineering–related programs at a large US university. The survey examined adoption patterns, task functions, verification practices, disclosure norms, and training needs. Undergraduates primarily used LLMs for tutoring and concept explanation (83%) and design ideation (67%), with limited adoption in coding (7%) and technical reasoning (41%). Verification practices were robust: 86% recalculated manually, 52% checked against standards, and only 4% reported nonverification, yielding a median of two methods per user. Ethical orientations favored conditional disclosure for major contributions (53%) and placed primary responsibility for errors on the human user (75%). Demand for formal training was high, especially among those with greater adoption, familiarity, and verification breadth. Results reveal a developmental gap between student practices, which emphasize low-risk learning and ideation, and research and faculty practices, which emphasize technically rigorous applications. The study underscores the need for curricular pathways that guide students from exploratory uses toward responsibly verified technical tasks within similar educational contexts. By linking a civil engineering–specific taxonomy of LLM functions with educational survey data, this work offers institutionally grounded empirical evidence on artificial intelligence (AI) literacy in civil engineering education and highlights directions for curriculum and assessment design.
This manuscript presents a descriptive study design and preliminary findings from an undergraduate engineering mechanics course conducted in Spring 2026, and details a reproducible survey instrument used to capture student AI usage patterns, attitudes, and verification practices, which are subsequently linked to academic performance metrics.
S. Geng, Helen Lallos-Harrell, Jiya Ashar et al.· arXiv.org· 0 citations
This study investigates the impact of large language model (LLM) usage, specifically ChatGPT, on student learning outcomes in programming education. The research adopts a mixed-methods approach, combining quantitative survey data from students and qualitative interviews with instructors. The study addresses three research questions: (1) the effect of LLM usage on undergraduate students' learning outcomes, (2) the influence of prompt engineering skills on this relationship, and (3) instructors' perceptions on these relationships. Quantitative data were collected from 159 students across two Saudi universities using a structured online survey with sections covering demographic information, LLM usage, self-reported programming understanding, and prompt engineering skills. Qualitative data were obtained through semi-structured interviews with programming instructors, covering LLM usage, prompt engineering skills, and their impact on student learning outcomes. The quantitative analysis utilized Partial Least Squares Structural Equation Modeling (PLS-SEM) to assess the measurement and structural models, including path coefficients, model explanatory power (R²), and predictive power (PLSpredict). Qualitative data were thematically analyzed using Atlas.ti to identify key themes related to instructor perspectives on the model. LLM usage positively impacts learning outcomes. While quantitative results did not show a significant moderating effect of prompt engineering skills, qualitative findings highlight its critical role in determining the positive effect of LLM usage on learning outcomes. The study emphasizes the importance of clear LLM usage policies and early prompt engineering training to promote meaningful engagement and maintain academic integrity in programming courses.
This study examined the electronic information resource (e-resource) utilization patterns, explicit academic motivations, and localized navigation barriers among undergraduate students at a private higher education institution (HEI) in the Philippines, to provide an empirical basis for optimizing digital library investments. Using a descriptive, retrospective document-analysis design, the study analyzed an existing institutional dataset consisting of 311 archived survey questionnaires collected during the 2025–2026 academic year across sixteen multidisciplinary academic programs. The findings revealed a pronounced discovery-use gap and collection polarization: the Philippine E-Journals (PEJ) platform accounted for the majority of student engagement (51.61%), well ahead of Google Scholar (30.87%) and EBSCOhost (7.40%). Undergraduate database interaction was largely transactional, driven primarily by compliance requirements for research papers (70.00%) and individual assignments (56.77%), and 64.52% of respondents rated the available collection as only “moderately relevant” to their research needs. A further, critical paradox emerged: although students reported high confidence in their baseline information technology (IT) skills, they encountered a substantial barrier when filtering search results, with the presence of irrelevant information identified as the most pervasive obstacle to effective database use. This difficulty was compounded by slow campus internet connectivity, while 56.13% of respondents had never received formal library training, and 90.00% left the program-adequacy item unanswered—a pattern consistent with students being unable to evaluate a service they had never experienced. The study concludes that substantial financial investment in commercial database subscriptions does not automatically translate into meaningful student engagement without structured instructional intervention. It is recommended that the institution establish a mandatory, credit-bearing Information Literacy Support (ILS) program, reallocate its e-resource budget based on actual usage data, formally embed targeted database tasks into course syllabi, and provide a dedicated, high-bandwidth Wi-Fi connection for the library’s electronic portals.
Roilingel P. Calilung· International Journal for Sc...· 0 citations
Context: Large Language Models (LLMs) such as ChatGPT are increasingly used in software engineering (SE) education, creating both opportunities and challenges that require systematic study for responsible curricular integration. Objective: This research seeks to build a validated framework for integrating LLMs into SE education through taxonomy development, empirical studies, and case analyses. This paper reports the first empirical step. Method: We mined 400 GitHub projects, analyzing README files and issue discussions to detect motivator and demotivator themes previously synthesized in our literature review [7]. Results: Key motivators included engagement and motivation (227 hits), software engineering process understanding (133 hits), and programming assistance & debugging support (97 hits). Prominent demotivators were plagiarism & IP concerns (385 hits), security, privacy & data integrity (87 hits), and over-reliance on AI in learning (39 hits). Demotivators such as challenges in evaluating learning outcomes and difficulty in curriculum redesign had no hits. Conclusion: These findings provide initial empirical validation of motivator/demotivator taxonomies, reveal research-practice gaps, and establish a basis for a comprehensive framework supporting responsible adoption of LLMs in SE education.
Maryam Khan· ACM Symposium on Applied Com...· 0 citations
Large language models (LLMs) are increasingly used by students and professionals in the architecture, engineering, and construction (AEC) sector for learning and information retrieval. However, their reliability in performing construction management (CM) knowledge tasks remains insufficiently characterized. This study introduces CMExamSet, a benchmark data set consisting of 689 multiple-choice questions compiled from four professional CM certification programs. The data set covers core CM domains, including project and program management, safety management, cost control, scheduling, contract administration, and related professional knowledge areas. Four contemporary LLMs were evaluated using a standardized zero-shot protocol with five repeated runs per question. Performance was assessed using accuracy, response consistency, subject-area analysis, and structured error annotation. Mean accuracy ranged from 79.2% to 90.0% across the four examinations, with high overall agreement observed in repeated runs. Performance varied systematically by domain: higher accuracy was observed in information retrieval-based domain tasks, such as engineering concepts, construction geomatics, and sustainability, whereas lower accuracy was observed in operationally oriented domains such as bidding and estimating, time and schedule management, and cost control. When errors occurred, conceptual misunderstandings were the most frequently observed error type across domains, indicating that incorrect interpretation or application of domain principles remained a common source of failure. Performance improvements in newer models were generally modest across subject areas. These findings provide empirical evidence on the capabilities and limitations of LLMs in CM knowledge assessment and underscore the importance of structured verification, domain grounding, and instructional guidance when integrating LLM tools into construction education and professional preparation.
Ruoxin Xiong, Yanyu Wang, Suat Gunhan et al.· Journal of Civil Engineering...· 0 citations
English-Medium Instruction (EMI) has become increasingly important in higher education, particularly in engineering programs where English is widely used as the medium of instruction. This study aimed to investigate undergraduate Electrical Engineering students’ perceptions of lecturers’ teaching methodologies in EMI classrooms, particularly in the Control Systems course at President University. A quantitative descriptive survey design was employed, involving 21 undergraduate students. Data were collected using a 20-item questionnaire measured on a 6-point Likert scale. Instrument reliability testing demonstrated excellent internal consistency (Cronbach’s alpha = 0.979). Data were analyzed using descriptive statistics, including frequencies, percentages, mean scores, and standard deviations. The findings revealed generally positive student perceptions of EMI teaching methodologies, with all questionnaire items obtaining mean scores above 4.00. Students reported that clear explanations, opportunities for interaction, and the use of supportive learning media enhanced their understanding and engagement in the learning process. Nevertheless, several students continued to experience challenges related to English language proficiency and technical terminology. These findings suggest that adaptive and student-centered teaching methodologies play an important role in maximizing learning effectiveness within EMI engineering classrooms.
Ida Bagus Putu Arkananta, Fiqih Nuraffi Rachmansyah, Ghania Rabbanni Hidayat et al.· Cerdika Jurnal Ilmiah Indone...· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduSep 14, 2026
The “HardFlow” algorithm could help generative AI models produce high-quality outputs that obey strict requirements when “pretty close” doesn’t cut it.
AI may appear weightless, but every model depends on physical infrastructure. To understand responsible AI, we need to look beyond algorithms and consider the entire lifecycle of the hardware behind them. The post Responsible AI Must Consider Its Afterlife appeared first on GPT-Lab.
