Navigating the Challenges of Large Language Models in Construction Education in South Africa: Balancing Innovation and Adoption

Nana Akua Asabea Gyadu-Asiedu, Clinton Aigbavboa, Moloisi Malesela David

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Digital technologies such as Large Language Models (LLMs) have transformed the construction industry, from design, planning, project management, execution, and building operations, with many processes now conducted remotely. However, integrating Large Language Models (LLMs) into construction education amidst rapid technological innovation presents significant challenges. The industry’s skills gap hampers productivity and innovation, while declining interest in construction-related university courses further underscores the need for innovative educational strategies. The study, therefore, explores the practical challenges of seamlessly integrating LLMs into construction education. Identifying such challenges emphasises the importance of understanding potential obstacles to ensure effective and responsible LLM adoption. This study explores these potential challenges through a comprehensive literature review and empirical findings from data collected via Google Forms from construction students, educators, and industry practitioners in Johannesburg, Gauteng. Convenience sampling was used, and data were analysed using inferential statistics such as the Kruskal-Wallis H test, an alternative to the one-way ANOVA, T-tests and Exploratory Factor Analysis. The findings reveal three clusters, which highlight the challenges as foundational challenges, context and ethical issues, and technical integration and usability issues. These findings make a case for the obstacles to LLM adoption in South African Construction Education. By investigating these challenges from industry and educational perspectives, the study concludes by recommending strategies to facilitate LLM integration, maximise benefits, and promote innovative, interactive, and skills-driven learning environments. It further emphasises the need to balance technological innovation with responsible integration to ensure positive outcomes for students, educators, and industry professionals.

Original languageEnglish
Title of host publicationProceedings of the Future Technologies Conference, FTC 2025, Volume 4
EditorsKohei Arai
PublisherSpringer Science and Business Media Deutschland GmbH
Pages227-244
Number of pages18
ISBN (Print)9783032079916
DOIs
Publication statusPublished - 2026
EventFuture Technologies Conference, FTC 2025 - Munich, Germany
Duration: 6 Nov 20257 Nov 2025

Publication series

NameLecture Notes in Networks and Systems
Volume1678 LNNS
ISSN (Print)2367-3370
ISSN (Electronic)2367-3389

Conference

ConferenceFuture Technologies Conference, FTC 2025
Country/TerritoryGermany
CityMunich
Period6/11/257/11/25

Keywords

  • Challenges
  • Construction education
  • Digital adoption
  • Large language models
  • Technological innovation

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Signal Processing
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Navigating the Challenges of Large Language Models in Construction Education in South Africa: Balancing Innovation and Adoption'. Together they form a unique fingerprint.

Cite this