About

Alex Albert is an Assistant Professor at the Department of Civil, Construction, and Environmental Engineering at North Carolina State University and serves as the Director of the North Carolina Construction Safety Laboratory (CSL). His research is focused on addressing safety challenges experienced in the construction industry. Through his leadership at CSL, he contributes to advancing knowledge and practices that improve construction safety management and worker protection. His work supports the development of safer construction environments by investigating factors that impact safety performance and by promoting best practices within the industry.

Research topics

• Computer Science
• Engineering
• Medicine
• Political Science
• Artificial Intelligence
• Economics
• Psychology
• Knowledge management
• Nursing
• Surgery
• Pedagogy
• Business
• Economic growth
• Simulation
• Forensic engineering
• Public relations
• Human–computer interaction
• Medical emergency
• Physical therapy

Selected publications

• Genders vs. dangers: Unpacking gendered disparities in hazard recognition, risk perception, and safety performance among construction workers

  Journal of Safety Research · 2026-04-27

  articleOpen access

  • The study examines gender as a perceptual bias in construction safety judgments. • Female workers showed higher hazard recognition performance than males. • No stimulus-based gender effect on hazard recognition was observed. • Male observers perceived higher risk when evaluating female workers. • Findings support bias-aware and gender-inclusive safety training practices in construction. Construction safety research has traditionally focused on individual, organizational, and environmental determinants of safety performance. However, little is known about how visual social cues, particularly gender, influence hazard recognition and safety risk perception during visual safety evaluations. This study examines whether gender functions as a perceptual bias in construction safety judgments using a reality–perception–expression framework. Seventy construction workers evaluated static images of real construction scenarios in which the gender of the depicted worker was systematically varied. Participants identified hazards, rated perceived safety risk, and responded to attitude statements related to gender and safety. Hazard recognition performance and safety risk perception were analyzed using linear mixed-effects models to account for repeated measures, with complementary nonparametric tests used where distributional assumptions were violated. Results showed that female participants demonstrated significantly higher hazard recognition performance than male participants (Mean = 35.80% vs. 28.45%; p = 0.007; Cohen’s d = 0.65). The gender of the worker depicted in the image did not significantly affect hazard recognition, indicating no stimulus-based gender effect. For safety risk perception, mixed-effects analysis revealed no overall gender differences; however, follow-up non-parametric analyses showed that male participants perceived higher risk when evaluating images depicting female workers (Mean = 0.14 vs. − 0.17; p = 0.015; Cohen’s d = 0.60), suggesting context-specific observer bias. Attitudinal results aligned with behavioral outcomes, with female participants more strongly rejecting male superiority in safety consciousness (p = 0.031) and endorsing women’s hazard recognition capability (p = 0.014). These findings demonstrate that gender-related differences in construction safety judgments are primarily observer-driven rather than stimulus-driven. While the use of static images enhances experimental control, it limits ecological validity, highlighting the need for future research using dynamic and immersive environments. Overall, this study advances understanding of the cognitive and social dimensions of construction safety and underscores the importance of bias-aware and gender-inclusive safety training and assessment practices.

  PublisherDOI

• Prompt to Protection: A Comparative Study of Multimodal LLMs in Construction Hazard Recognition

  IEEE Access · 2026-01-01

  articleOpen accessSenior author

  The recent emergence of multimodal large language models (LLMs) has introduced new opportunities for improving visual hazard recognition on construction sites. However, despite growing interest in their applications, there has been limited investigation into how different LLMs perform in safety-critical visual tasks within the construction domain. To address this gap, this study conducts a comparative evaluation of five state-of-the-art LLMs: GPT-4o, GPT-5, GPT-4.1, Claude 4.1 Opus, and Gemini 2.5 Pro, to assess their ability to identify potential hazards from real-world construction images. Each model was tested under three prompting strategies: zero-shot, few-shot, and chain-of-thought (CoT). Quantitative analysis was performed using precision, recall, and F1-score metrics across all conditions. Results reveal that prompting strategy significantly influenced hazard recognition performance. CoT prompting consistently produced the highest accuracy across models, with GPT-5 and GPT-4.1 achieving superior scores in most settings. The results suggest that structured prompt design can elevate LLM performance to significant levels, offering a cost-effective pathway for small firms and contractors to improve hazard recognition. Furthermore, LMM outputs can be repurposed as explanatory content for safety training and toolbox talks, supporting long-term safety culture development.

  PublisherDOI

• Comparing training delivery methods: Impact on learning outcomes and engagement among construction workers

  Safety Science · 2025-04-08 · 4 citations

  articleOpen access

  • This study compares five safety training methods using a multi-arm field experiment on construction sites. • Training methods were evaluated for their effect on both learning outcomes and worker engagement levels. • Data were collected from 591 workers using pre- and post-training surveys on safety knowledge and engagement. • Pre-recorded video improved learning at low cost; hands-on interactive lectures boosted both learning and engagement. • Teaching an energy-based hazard definition led to a 26% improvement in workers’ hazard recognition scores. Effective safety training is crucial for enhancing workers’ safety awareness and promoting safer behaviors, yet delivering such training within time and budget constraints remains a challenge. This study assessed the effectiveness of five safety training delivery methods—pre-recorded video, lecture, interactive lecture, flipped lecture (pre-recorded video followed by a hands-on activity after two weeks), and interactive lecture with hands-on activity—on engagement levels and short-term learning outcomes, namely hazard recognition skill, high-energy hazard recognition skill (hazards with the potential to cause serious injuries or fatalities), perception of high-energy hazards, and risk tolerance. A multi-arm parallel-group field experiment was conducted, with data collected through pre- and post-training surveys. The results revealed that while engagement levels increased linearly as the learner-centeredness of the training increased, improvements in hazard recognition skills followed a parabolic trend. Additionally, trainees in all groups except the interactive lecture group perceived high-energy hazards as riskier after the training but risk tolerance did not decrease in any training groups. These findings provide specific guidance on selecting optimal safety training delivery method as it relates to training objectives and resource constraints. These results may help practitioners to select the most appropriate method of safety training delivery based on their specific aims and available resources.

  PublisherDOI

• Developing a Fit-for-Purpose Best Practice Knowledge Handbook Using Generative AI

  2025-07-31

  articleOpen access

  The Construction Industry Institute (CII) has produced extensive research shown to deliver cost, schedule, and safety benefits within the construction industry.However, low visibility and limited accessibility to the research have hindered its widespread adoption.This paper introduces a generative artificial intelligence (GenAI) framework designed to enhance the accessibility and utilization of CII research for industry practitioners.The proposed framework employs a Retrieval-Augmented Generation (RAG) approach by integrating CII best practice (BP) research reports into a large language model (LLM) to identify relevant insights.First, a hybrid method combines qualitative analysis with GenAI-driven questionanswering to extract critical findings ("golden nuggets") from each BP report.These nuggets are then used to generate detailed action items (DAIs) by LLM.The validation process involves using an LLM judge method complemented by subject matter experts (SMEs) review.Lastly, an executive summary and frequently asked questions (FAQs) were generated for each BP by feeding GNs and DAIs information to LLM.As a result, a comprehensive BP knowledge handbook was generated.This handbook showcases the potential of GenAI in construction knowledge extraction in a passive way.Additionally, this study will also present a prototype where users can interact with the GenAI chatbot in an active way for knowledge harvesting and communication.This study contributes to advancing human-AI interaction in construction knowledge management, offering a scalable and user-friendly solution for bridging the gap between research and practice.

  PublisherOA PDFDOI

• A Novel Edge Computing Framework for Construction Nail Detection under Conditions of Constrained Computing Resources

  2025-12-11

  articleCorresponding

  Construction housekeeping promotes a safe working environment but often lacks an effective approach for monitoring and maintenance. Recent research has used artificial intelligence (AI) techniques to detect construction objects automatically to facilitate the work of safety managers. However, none of the research has considered implementing real-time AI applications for automatic construction housekeeping monitoring. We propose a framework that integrates edge computing and a computer vision model to detect boards with nails that may be scattered throughout construction sites. First, we trained a MobileNet machine learning (ML) model to identify boards with or without nails. Then, we quantized the model using TensorFlow Lite to allow the model’s optimal deployment in edge devices. Lastly, we assembled an edge device module based on Raspberry Pi and embedded the ML model to realize real-time offline housekeeping monitoring. The experimental results show great promise for both lab settings and in practice at construction sites. The proposed framework can facilitate AI applications in various construction fields under computing resource-constrained conditions.

  PublisherDOI

• How Reliable Are Large Language Models? Zero-Shot Detection of Construction Hazards

  Proceedings of the ... ISARC · 2025-07-27

  articleOpen access

  The construction industry persistently underperforms in hazard recognition, often leading to severe workplace injuries due to unrecognized hazards.With the recent advancements in Artificial Intelligence (AI) and the emergence of Large Language Models (LLM), the construction sector has begun exploring these technologies for various applications.However, a systematic comparison of popular LLMs to evaluate their effectiveness in identifying construction hazards remains unexplored.Additionally, previous studies have primarily focused on assessing LLMs using textual input and output, leaving their performance with visual inputs underexplored.This study addresses this gap by systematically assessing and comparing the hazard recognition performance of five widely used LLMs using construction case images.The findings establish a baseline standard for LLMs in construction hazard identification through zero-shot learning and reveal that LLMs do not perform significantly well in this context.Additionally, the study provides valuable insights into the reliability and potential applications of LLMs for enhancing hazard recognition in the construction industry.

  PublisherOA PDFDOI

• Toward the Development of a Fit-for-Purpose Handbook for the Upstream, Midstream, and Mining (UMM) Sector

  2024-03-18 · 2 citations

  articleSenior authorCorresponding

  Projects in the upstream, midstream, and mining (UMM) sectors present unique complexity. The Construction Industry Institute’s (CII) Research Team 398 aimed at developing the criteria that define project complexity in the UMM sector, a matrix that describes the different levels of complexity and assign tools from the CII database that can be used to manage complexity. The eventual goal is to create a fit-for-purpose handbook for the UMM sector. This project is divided into phases; the first phase is to develop a complexity matrix for the UMM sector CII companies with more mature project management organizations. The team used convenience sampling and conducted interviews with owners and contractors for this purpose. The methodology also included reviewing literature from research into complexity, the characteristics of the UMM sector and incorporating the findings of other complexity models developed within and outside of CII. The result was a complexity matrix that includes factors relevant to all the three sectors. The vision for the next phase includes validating the current complexity matrix and identifying key information (“golden nuggets”) for addressing complexity.

  PublisherDOI

• Validation of mobile phone use recall in the multinational MOBI-kids study

  Utrecht University Repository (Utrecht University) · 2024-10-01

  articleOpen access

  Potential differential and non-differential recall error in mobile phone use (MPU) in the multinational MOBI-Kids case-control study were evaluated. We compared self-reported MPU with network operator billing record data up to 3 months, 1 year, and 2 years before the interview date from 702 subjects aged between 10 and 24 years in eight countries. Spearman rank correlations, Kappa coefficients and geometric mean ratios (GMRs) were used. No material differences in MPU recall estimates between cases and controls were observed. The Spearman rank correlation coefficients between self-reported and recorded MPU in the most recent 3 months were 0.57 and 0.59 for call number and for call duration, respectively. The number of calls was on average underestimated by the participants (GMR = 0.69), while the duration of calls was overestimated (GMR = 1.59). Country, years since start of using a mobile phone, age at time of interview, and sex did not appear to influence recall accuracy for either call number or call duration. A trend in recall error was seen with level of self-reported MPU, with underestimation of use at lower levels and overestimation of use at higher levels for both number and duration of calls. Although both systematic and random errors in self-reported MPU among participants were observed, there was no evidence of differential recall error between cases and controls. Nonetheless, these sources of exposure measurement error warrant consideration in interpretation of the MOBI-Kids case-control study results on the association between children's use of mobile phones and potential brain cancer risk.

  Publisher

• Biggest Challenges Facing the Construction Industry

  2024-03-18 · 16 citations

  articleSenior authorCorresponding

  The construction industry is one of the main pillars of the economy. It plays a vital role in the country’s economy by generating employment, upgrading current infrastructure, and boosting overall economic activities. In the United States, the economic contribution of the construction sector to the gross domestic product is nearly 4%. Nevertheless, the construction industry has not reached the desirable levels of overall performance in terms of delivering projects on time and within budget while maintaining high levels of safety, quality, and sustainability. Consequently, it is critical to pinpoint the industry’s challenges to provide practical solutions. In this study, 25 experienced construction professionals were interviewed to understand better the emerging and persistent challenges of the construction industry. The study findings suggest that the most significant and pressing challenges facing the construction industry are (1) the shortage of skilled labor, (2) supply chain disruptions along with material and labor cost volatility especially after the COVID-19 pandemic, and (3) the slow integration, high cost, and maturity of new emerging technologies (e.g., 3D printing, virtual and augmented reality). The study findings will help industry leaders, governmental agencies, and construction researchers develop solutions that would improve the overall performance of the construction industry.

  PublisherDOI

• Meet2mitigate: An Llm-Powered Framework for Real-Time Issue Identification and Mitigation from Construction Meeting Discourse

  SSRN Electronic Journal · 2024-01-01

  preprintOpen access
  PublisherDOI

Frequent coauthors

• S M Jamil Uddin

  Florida Gulf Coast University

  84 shared

• Abdullah Alsharef

  King Saud University

  81 shared

• Anto Ovid

  North Carolina State University

  54 shared

• Gemma Castaño‐Vinyals

  Centro de Investigación Biomédica en Red de Epidemiología y Salud Pública

  48 shared

• Siddharth Bhandari

  University of Nebraska–Lincoln

  29 shared

• Chelsea E. Langer

  New Mexico Department of Health

  27 shared

• Elisabeth Cardis

  Barcelona Institute for Global Health

  27 shared

• Brigitte Lacour

  Inserm

  24 shared

Labs

• North Carolina Construction Safety LaboratoryPI

  Addressing safety challenges experienced in the construction industry

Education

• Ph.D., Civil Engineering

  University of XYZ

  2005

• M.S., Environmental Engineering

  University of ABC

  2002

• B.S., Civil Engineering

  XYZ State University

  1999

Awards & honors

• 2017 Best Paper Award, Journal of Construction Engineering a…
• 2016 Outstanding Reviewer, Journal of Construction and Engin…
• 2015 Outstanding Reviewer, Journal of Construction and Engin…
• 2015 New Scholar, Construction Industry Institute (CII)
• 2014 Best Paper (Second), Construction Research Congress, (C…