People are inextricably linked with the buildings and infrastructure that cumulatively constitute the modern built environment (Edwards et al., 2017). They must live and work together in harmony with the built environment while also maintaining equilibrium with the natural environment (Edwards et al., 1996; Seidu et al., 2025). This objective is core to governments globally who must generate wealth for powering renewed industrial growth (Raihan et al., 2023). To achieve such a delicate multi-dimensional balance, advanced digitalisation and artificial intelligence (AI), under the umbrella of Industry 4.0, are increasingly seen as a panacea to environmental sustainability, public safety and business performance issues (Newman et al., 2021). Indeed, a wave of exciting digital and AI technologies (including generative AI) are now commercially available, with governments and industry across the global clambering to capitalise upon this revolution (Darko et al., 2020). Applications are myriad and include autonomous vehicles (Edwards et al., 2022), advanced automation (Yang et al., 2023) and blockchain technology (Bayramova et al., 2021) – all of which seek to create a better built environment for humanity. That said, this latter-day gold rush is primarily powered by the neoliberal economic environment that drives mass consumption and concomitant anthropogenic emissions (Roberts and Edwards, 2022). Evidence of this claim is substantiated by reports of $79bn being generated for AI founders (Carson et al., 2025) and the onset of the world's first trillionaire (Jamali et al., 2025). The allure of gold is most, and as Plato once stated, “All the gold upon the earth and all the gold beneath it do not compensate for a lack of virtue.” Ironically, real riches are obtained from the pursuit of knowledge, which is far more precious and enduring than gold ever will be – otherwise this philosophical quote would have been forgotten many years ago.
AI-driven solutions are, however, not without their issues. For example, AI-driven data centres, used to drive sustainable development, remain mass consumers of energy derived from fossil fuels (Sun and Lee, 2006). Moreover, literature used as a basis for driving innovation and education is now subject to epistemic contamination (Athaluri et al., 2023). Akin to forever chemicals (Duru, 2023), such contamination has a multiplicative impact because it infects the whole knowledge supply chain from research paper writing through to student learning using polluted materials. Data hallucination is another factor that contaminates literature due to generative AI's innate ability to manufacture false narratives (Kacena et al., 2024) and/or draw from material that may be subject to logical fallacies. More worrying is the negative impact upon humanity and our cognitive functionality (Walther, 2025). AI would appear well suited to deterministic problems where the math and science are premised upon finite algorithms (such as gravity) that can be predicted precisely. They are also highly valuable for analysing complex big data sets expediently (Ghosh et al., 2020). For more stochastic problems that cannot be predicted definitively, AI is not as accurate because it is not sentient. Perhaps AI should be renamed Preprogrammed Algorithms (PA), since logic trained with the PA is only as good as the people creating the code?
These and other cautionary tales provide a reality check on the current status quo of AI – which undoubtedly has a huge transformative influence upon social, economic and political policy. AI limitations must not discourage scientists from testing the boundaries of implementation to engender smart infrastructure and safety innovations. New knowledge is founded upon curiosity, scientific endeavour and a rich discourse to generate thesis and antithesis. In the absence of competing debate, discovery and progress would not be made. Perhaps the lesson is that humanity must maintain control of the technology and not be enslaved by it (cf. Edwards et al., 2017). AI certainly has its place and is unlikely to be disinvented but must be used judiciously and with integrity.
On this note, it is good to read 13 papers generated for this issue on “Smart infrastructure and safety innovation: AI-driven risk management and resilience strategies.” Such work expands the frontiers of knowledge to generate new insight and perspective regarding the challenges facing humanity and how we may overcome them. More innovative work of this nature is needed because optimal solutions presented will invariably be adopted in practice for the benefit of humanity. An exciting future lies ahead for the construction industry and academics within higher education institutions who are actively tackling the challenges posed to mitigate the risks.
Sudhakaran et al. (2025) examine how extended reality (XR) technologies can enhance safety for vulnerable road users such as pedestrians and cyclists. In their study, Sudhakaran et al. (2024) conducted a systematic review of 80 articles to understand XR's potential in transportation safety. They found that XR shows promise beyond just pedestrians and cyclists, and combining eye-tracking with virtual reality could be particularly effective. The findings suggest XR can help urban planners and transportation authorities create better safety solutions. This technology offers new ways to train vulnerable road users and design safer urban environments. This study fills a critical gap by demonstrating how emerging technologies can address road safety challenges without compromising people's safety.
Sarhan et al. (2025) investigate occupational stress among construction workers and professionals in the UK. Stress, anxiety and depression are the second leading cause of work-related illness in UK construction, yet few studies examine their sources. The authors used interviews with construction professionals and a focus group with industry experts to identify stress factors. They identified 7 primary stressors and 35 secondary influencing factors in construction projects. Workflow interruptions emerged as the most significant stressor. The study reveals the limitations of traditional project planning methods, such as the critical path method. It recommends inclusive and collaborative planning as a strategy to prevent stress. The findings can help project managers address stressors early in construction projects. The study suggests that improving workflow and collaboration could enhance worker health, well-being and productivity. It also recommends incorporating workflow management into official stress prevention standards.
The subsequent study aims to inform designers, architects and planners in creating more accessible and inclusive public spaces. In this study, Carnemolla et al. (2025) examine how wheelchair users experience public accessible bathrooms and what they think about current design standards. While public bathrooms are highly regulated for accessibility, little research exists on wheelchair users' actual experiences. The authors conducted in-depth interviews with users of powered and manual wheelchairs to gain insight into their perspectives. They explored how participants use accessible bathrooms based on their mobility needs, disability, support levels and the type of wheelchair they use. The findings reveal four main themes: safety, hygiene, planning and/or avoidance and privacy and dignity. Many wheelchair users spend considerable effort planning bathroom use or avoid public bathrooms entirely. This study emphasises the importance of ongoing maintenance and regular cleaning for ensuring accessibility and safety. However, the maintenance aspects are not currently captured in design regulations. This demonstrates that both design and maintenance have a significant impact on the health, well-being and social participation of people with disabilities. The findings can help improve bathroom design regulations to include maintenance guidance.
The following study provides a user-friendly tool for decision-making processes involving public-private partnerships. In this research, Tajani et al. (2025) present a decision-support model for managing limited resources in enhancing cultural heritage and promoting the sustainable development of small towns. The model uses a multi-criteria method that combines linear planning with performance indicators. These indicators capture the complexity of a territory's cultural identity and existing cultural-historical assets. The authors tested the model in a case study in a municipality in southern Italy. The findings demonstrate that the model enables public and private operators to make informed decisions about which preservation and restoration projects to fund. It allows them to consider the economic, social and environmental impacts of their choices. This research contributes to the sustainable management of cultural heritage and comprehensive territorial development.
Adepu et al. (2025) investigate how the COVID-19 pandemic led to cost overruns in construction projects. The pandemic created multiple challenges to the industry, including material shortages, project delays, labour shortages, increased costs and funding difficulties. While many studies have examined the impact of COVID-19 on construction budgets, few have accurately predicted the magnitude of this impact. The authors developed a predictive tool using ordinal logistic regression methods. This tool was developed based on an online survey conducted with construction industry experts to understand the factors contributing to excessive costs during the pandemic. The findings reveal that smaller enterprises and contractor-focused organisations faced greater challenges than medium-to-large companies and consultancy or owner-type entities. The research highlights the increased risk of higher project costs during health crises or similar events. It emphasises the need for stronger risk management strategies to improve project outcomes. By considering demographic factors, the study provides policymakers with insights to customise interventions for the construction industry.
Jouma et al. (2025) investigate the daily performance of a microgrid system that combines photovoltaic panels and wind turbines connected to the main electricity grid. The microgrid serves a residential area with 20 houses. The study examines a daily operational strategy that allows the microgrid to buy and sell energy with the main grid. Smart meters provide suppliers with daily consumption patterns, which can be improved through demand-side management (DSM). The researchers used daily operational cost, CO₂ emissions and other measures to evaluate system performance. They employed a grey wolf optimiser to minimise costs, including energy procurement, emission costs and revenue from energy sales. The results for winter and summer days indicate that DSM has significantly improved both economic and environmental performance. With DSM, winter day costs were −26.93 (/kWh) and summer day costs were 10.59 (/kWh). Without DSM, costs were −25.42 (/kWh) in winter and 14.95 (/kWh) in summer. Unlike previous research that focused on long-term operations, this study examines short-term (24-h) microgrid operation. It addresses critical challenges of selling and buying energy while considering environmental costs. The study engages consumers through smart meters for demand-side management, whereas previous studies have primarily focused on supply-side management.
In the subsequent study, Alsemari and Ramegowda (2025) provide evidence to encourage digitalisation and reduce CO₂ emissions in Iraq's oil and gas sector through the Basra Oil Company (BOC). The oil and gas sector is Iraq's primary economic resource, and projects require substantial expenditures due to the unique facility requirements. BIM technology has not yet been adopted in Iraq despite its potential benefits. Conducting a literature review, the authors developed a hypothesis that adopting BIM instead of traditional 2D approaches could help BOC overcome project management constraints. They designed a web-based questionnaire and distributed it to BOC professionals with diverse engineering backgrounds to test their readiness and acceptance of BIM. Results showed that respondents accepted BIM as an alternative to current 2D traditional approaches. They identified and ranked BIM implementation barriers based on their criticality. Findings revealed that the company's top management and policies are the most critical factors for BIM adoption. Cost was not seen as a critical barrier to implementing BIM in this sector. The research demonstrates that BIM is valuable for managing oil and gas projects in Iraq.
The construction industry has a complex and evolving supply chain with multiple stakeholders and processes. Enhancing supply chain resilience is crucial for industry growth and competitiveness in the face of increasing global challenges. To address this gap, Singh et al. (2025) develop a framework for identifying and classifying critical success factors (CSFs) for supply chain resilience in sustainable construction. Conducting a comprehensive literature review, a sequential mixed-methods approach and expert consultations, the authors identified the potential CSFs. The fuzzy DEMATEL method was applied to classify CSFs into cause-and-effect groups. Interpretive structural modelling (ISM) established a hierarchical framework to analyse interrelationships among CSFs. The study identified 17 finalised CSFs, with 12 causal factors and 5 effect factors across 5 hierarchical levels. Safety measures and guidelines have the most significant potential to drive change. The availability of standardised materials and equipment, as well as skilled human resources, depends highly on causal CSFs. The findings provide detailed insights to bridge the gap between unsustainability and sustainability by enhancing the resilience of construction supply chains. The framework provides practical guidance on allocating resources to strengthen construction supply chain resilience.
Achieving sustainable construction is highly challenging in developing economies due to inadequate technical support systems. To this end, Moyo et al. (2025) develop a technical support system (TSS) framework for sustainable construction indicators in Zimbabwe. Their study aims to address this gap by creating a comprehensive framework for Zimbabwe as a developing country. The authors adopted a post-positivist philosophical approach and deductive reasoning to test relevant theory. Construction professionals from consultancy firms, construction companies, government bodies and academic institutions participated in an online questionnaire survey. The fuzzy synthetic evaluation (FSE) revealed a framework with six critical technical support subgroups ranked by importance. These include innovation for construction sustainability, adequate expenditure on sustainability, skills training support, project economic assessment and governance support, circularity and environmental technical support, climate change literacy and supplier assessment support and decent work support. This framework can enhance existing sustainable construction initiatives by supporting related policy initiatives. The findings offer practical guidance for developing economies facing similar challenges in implementing sustainable construction practices.
Housing energy consumption significantly contributes to climate change, and promoting energy-efficient products can reduce this consumption. In this aspect, Du et al. (2025) investigate how social influences affect people's decisions to adopt energy-efficient products at different price levels. While previous research has identified social influences such as peer effects and social norms, these influences have not been quantified or compared with one another. The authors conducted two stated choice experiments in Wuhan, China, examining low-cost and high-cost scenarios. Appliance packages (fridges and washing machines) represented the low-cost scenario, while heating and cooling systems represented the high-cost scenario. Social influences were evaluated across three aspects: positive versus negative information, physical versus online social networks and peer effects versus social norms. The findings reveal that social influences have a greater and broader impact in low-cost scenarios than in high-cost scenarios. Negative information decreases the adoption of low-cost, energy-efficient products, while positive information boosts the adoption of high-cost products. People value information from those they know personally and are more influenced by physical social networks than online ones. This study presents a comprehensive framework for investigating and comparing social influences across various processes, paths and information types. The findings offer practical implications for policymakers to accelerate energy transition in the built environment.
High-rise construction involves unique characteristics, complicated design, hazardous activities and complex work environments that create significant risks worldwide. In India, construction safety management is crucial due to the substantial occupational risks and hazards present in workplaces. These occupational hazards lead to accidents that have a severe impact on human health and result in significant financial losses. In their study, Tripathi and Mittal (2025) address occupational safety risks in high-rise construction projects in India. The authors developed a hybrid approach combining the risk assessment method (RAM) and the technique for order of preference by similarity to ideal solution (TOPSIS) to detect and evaluate occupational risks. Using a questionnaire survey, the study identified six types of construction activities and ten corresponding risks. Based on risk score calculations, roof work activities, finishing work and mechanical, electrical and plumbing (MEP) work were identified as the most hazardous construction activities. Among the ten risks identified, workers falling from height is the most prominent risk across the majority of activities. Other major risks include fire, electrical accidents and being struck by falling objects. By integrating theoretical insights with practical applications, the study enhances occupational safety and aims to reduce construction site accidents, contributing to both academic knowledge and industry practices.
In the subsequent study, Mohandes et al. (2025) address the critical issue of sewer overflows (SO) and their role in achieving smart and sustainable urban drainage in cities. Well-functioning sewer systems are crucial for urban sustainability; however, sewer overflows have detrimental impacts on the environment and public health. This study aims to bridge significant research gaps by investigating the root causes of sewer overflow incidents and understanding their broader ecological consequences. The authors developed the multi-phase causal inference fuzzy-based framework (MCIF) to address these gaps. MCIF integrates the fuzzy Delphi technique, fuzzy DEMATEL method, fuzzy TOPSIS technique and expert interviews. Drawing on expertise from developed countries, the framework systematically identifies and prioritises sewer overflow causes, explores causal relationships, prioritises environmental impacts and compiles mitigation strategies. The findings effectively identify and prioritise causal factors behind sewer overflow incidents, highlighting their relative significance. The study reveals intricate causal relationships among key factors, including blockages, flow velocity, infiltration and inflow, under-designed pipe diameter and pipe deformation, holes or collapse. These insights offer a profound understanding of the complex factors that contribute to sewer overflows.
Most safety research focuses on drivers, pedestrians and vehicles, with little attention given to highway traffic officers (HTOs). Accordingly, Bortey et al. (2025) develop a prediction model for incidents involving HTOs. This study aims to enable highway safety authorities to predict exclusive incidents such as incursions and environmental hazards, respond effectively to safety risk scenarios and implement timely precautions to minimise HTO incidents. The authors used data from a highway incident database and applied supervised machine learning methods. Three algorithms were employed – support vector machine (SVM), random forests (RF) and Naïve Bayes (NB) – and their performances were comparatively analysed. Three data balancing algorithms were also applied to handle class imbalance challenges. The methodology consisted of five sequential phases: data collection, data preprocessing, model selection, data balancing and model evaluation. Their findings indicate that an SVM with a polynomial kernel, combined with the Synthetic Minority Over-sampling Technique (SMOTE) algorithm, is the best model for predicting various incidents. The Random Under-sampling (RU) algorithm was the most inefficient in improving model accuracy. Weather and/or visibility, age range and location emerged as the most significant factors in predicting highway incidents. This model provides evidence-based information to train HTOs on impending risks, equipping workers with awareness, anticipation and flexibility to handle potential hazards.
