Job demands-resources (JD-R) theory is used to describe the relationships between burnout in terms of job demands and work engagement in terms of job resources (Lazarova et al., 2010). (Xia et al. 2020) proposed that all job characteristics can be classified into job demands and job resources according to the JD-R theory. Job demands relating to specific contextual factors refer to “those physical, psychological, social, or organizational aspects of the job that necessitate sustained physical and psychological effort and are consequently associated with specific physiological and psychological costs” (Xia et al., 2020). Job resources relating to the enhancement of work engagement are defined as “those physical, psychological, social, or organizational aspects of the job that are functional in achieving work goals, reduce job demands and the associated physiological and psychological costs, or stimulate personal growth, learning, and development” (Cheung et al., 2021; Xia et al., 2020). In a particular context, the interactions between work demands and job resources are assumed to be supported by health impairment and motivational processes in JD-R theory (Yaris et al., 2020). In particular, the health impairment process is designed to elucidate the manner in which job demands deplete employees' energy and result in burnout, whereas the motivational process demonstrates the mechanisms by which job resources promote employees' positive behaviors to achieve relevant goals and facilitate work engagement (Cheung et al., 2021; Conchie et al., 2013; Yaris et al., 2020). In the context of Construction 5.0, workers encounter increased safety challenges that require them to adapt to new technologies and performance expectations. The cognitive and time investments needed to meet these evolving safety protocols surpass those of traditional practices, making it crucial to fulfill safety demands effectively. The ability to meet these safety demands significantly influences the safety practices of construction workers. Psychosocial support, safety participation, and leadership safety behavior can enhance safety-related job resources by providing emotional backing, improving safety perceptions, and addressing organizational issues. Such support can motivate workers to adopt safer practices, increasing their engagement in safety-related tasks and ultimately enhancing safety outcomes in Construction 5.0. Therefore, the motivational processes highlighted in JD-R theory offer a valuable framework for understanding how psychosocial support, safety participation, and leadership safety behavior impact safety practices in this modern construction environment.

Safety practice in Construction 5.0 embodies a human-centric approach that proactively integrates advanced technologies with worker behavior to optimize occupational wellbeing (Gambatese John et al., 2017; Osborne et al., 2024). Emerging as an evolution from conventional safety management, this concept transcends mere safety performance evaluation by emphasizing construction workers' real safety behaviors and actual safety outcomes in technology-intensive workspaces (Hwang and Soh, 2013). Organizations invest in tools such as AI-driven training programs, IoT sensors for hazard detection, and ergonomic exoskeletons, while workers adapt to interfaces like augmented reality (AR) safety guides and robotic coordination systems (Han et al., 2020; Wijayarathne et al., 2024). In dynamic Construction 5.0 settings, including hybrid workspaces where humans and machines operate collaboratively, safety outcomes depend not only on resources but also on effective interactions between workers and technology (Liang et al., 2021a; Zhang et al., 2016). For instance, wearable devices that adjust workflows based on fatigue data demonstrate how proactive behaviors and technological tools jointly mitigate risks (Bayram, 2020; Feng et al., 2014). At the organizational level, safety practices involve simulating protocols through digital twin models to predict and prevent hazards (Yan et al., 2025). At the individual level, workers engage with adaptive tools such as smart helmets that provide real-time feedback on posture or equipment use (Ellis and Lee, 2006). Ultimately, more effective safety practices in Construction 5.0 are achieved through workers' exhibition of safer behaviors that generate positive safety outcomes, supported by advanced technological tools.

Safety practices in Construction 5.0 focus on achieving effective safety outcomes through the safe and compliant behaviors of construction workers. It ensures that safety measures are both practical and impactful, helping to reduce accidents in the industry. Various stakeholders (including owners, designers, contractors, and vendors) collaborate to implement these safety measures (Choe et al., 2020). The objective is to enable organizations and workers to effectively utilize safety resources to meet their safety goals (Pritchard et al., 2008). Improvements in safety practices, supported by real-time monitoring and predictive analytics, can significantly lower the risk of project failures and minimize time and cost overruns related to safety management (Hwang and Soh, 2013; Krajangsri and Pongpeng, 2017; Nasirzadeh and Nojedehi, 2013). While psychosocial factors, safety climate, and organizational behaviors remain critical in Construction 5.0, they are increasingly influenced by technological interfaces (Mitropoulos and Memarian, 2012). Issues such as absenteeism and presenteeism are linked to psychosocial factors like social support and the physical work environment, which can negatively impact safety outcomes (Abdullah et al., 2015; Bayram, 2020; Johns, 2010). The safety climate in Construction 5.0 is shaped by shared perceptions of safety policies and practices, as well as trust in automated systems and the transparency of AI-generated safety protocols (Ohueri et al., 2024; Xia et al., 2021). This evolving climate indirectly influences workers' safety behaviors, affecting their compliance with hybrid workflows that combine human decisions with machine recommendations (Barling et al., 2002; Bonham et al., 2017; Cooper and Phillips, 2004). Organizational safety behaviors in Construction 5.0 focus on upskilling workers to safely interact with automated machines, fostering psychological safety in data-driven decision-making, and encouraging proactive reporting of system vulnerabilities (Wang et al., 2021; Wijayarathne et al., 2024). These initiatives create a secure human-technology ecosystem and facilitate effective safety practices through adaptive governance frameworks (Ford and Tetrick, 2008).

Psychosocial support is defined as a form of support characterized by friendship and caring that extends beyond work (Higgins et al., 2010). In Construction 5.0 environments where human-robot collaboration becomes increasingly common, psychosocial support emphasizes the establishment of emotional support and trust through hybrid human-digital interfaces, which can provide support in the areas of care and friendship while maintaining technological synergy (James, 2000; Zhuang et al., 2013). Emotional support in smart construction sites requires demonstrating courtesy and regard for construction workers through both interpersonal interactions and digitally-mediated communication channels, which reduces technostress levels in technology-intensive work environments (Tews et al., 2020). Contemporary trust-building extends beyond human relationships to encompass triadic interactions among workers, collaborative robots, and AI systems, facilitated by algorithmically transparent processes that enhance human-machine team cohesion (Chua et al., 2008). Mentoring functions and psychosocial factors remain crucial in Construction 5.0 contexts. Individuals receiving psychosocial support through AR training systems exhibit higher professional optimism, enhancing their resilience in adapting to digital-physical work ecosystems while reducing automation-related stress (Qin and Bulbul, 2023). Their career outcomes now include digital skill satisfaction and emotional investment in sustainable construction practices (Higgins et al., 2010). Advanced implementations integrate IoT-embedded social support systems that perform continuous behavioral monitoring, enabling predictive wellbeing interventions to reduce burnout and real-time hazard detection to prevent accidents (Huang et al., 2019; Sobeih et al., 2009). While traditional demographic variables remain relevant, Construction 5.0 introduces digital literacy as a critical demographic determinant influencing psychosocial support accessibility (James, 2000). The reconceptualized psychosocial perspective advocates hybrid supervisory models that synthesize AI-powered performance analytics with human-centered leadership principles (Callari et al., 2024). These models leverage collaborative digital twins as immersive communication platforms to optimize team dynamics, effectively bridging technological capabilities with psychosocial needs (Du et al., 2020; Ingelgård et al., 1996).

Safety management in Construction 5.0 establishes multidimensional job requirements that integrate intelligent monitoring systems with human-centered safety protocols (Chauhan et al., 2024; Chen and Chan Isabelle, 2024; Ohueri et al., 2024). Construction workers' safety practices emerge as the critical output of their safety input, mediated through the expenditure of safety-related resources (Bayram, 2020; Dai et al., 2009). In this context, psychosocial support plays a crucial role in technology-intensive workspace by enhancing workers' social connections and self-efficacy—key emotional resources needed for effective human-machine collaboration Dollard et al., 2000; Higgins et al., 2010). According to JD-R theory, when construction workers receive adequate psychosocial support, they are better equipped to manage the cognitive demands of hybrid physical-digital safety operations (Xia et al., 2023). This support helps maintain their cognitive engagement with safety objectives, leading to improved compliance with adaptive safety regulations (Higgins et al., 2010; James, 2000; Lee et al., 2020). Therefore, psychosocial support serves as a vital resource within Construction 5.0, compensating for the psychological effort needed to sustain safety vigilance in technology-driven environments (Ohueri et al., 2024; Pirzadeh et al., 2022). It also promotes positive safety outcomes through enhanced peer-to-peer knowledge sharing (Higgins et al., 2010; Lee et al., 2020). This mechanism suggests that psychosocial support interventions can significantly enhance the effectiveness of safety investments in Construction 5.0 and foster sustained safety behaviors and practices that exceed basic compliance. Thus, the first hypothesis is proposed as follows:

Safety participation is defined as behaviors that do not directly contribute to an individual's personal safety but that do help to develop an environment that supports safety (Xia et al., 2018a). In the context of Construction 5.0, this concept expands to include collaborative behaviors between humans and technology, where workers engage with digital tools such as digital twins and safety systems (Wijayarathne et al., 2024). Modern forms of safety participation involve activities like AR-enhanced safety drills, collaborative problem-solving on digital safety platforms, and proactive reporting of issues within building information modeling (BIM) systems (Hussain et al., 2024; Shekargoftar et al., 2022). As a critical element of employee safety performance behaviors, safety participation in Construction 5.0 not only sustains traditional safety climate but also optimizes human-cyber-physical system interoperability in smart construction ecosystems (Ohueri et al., 2024). This integration allows for the identification of system dysfunctions and promotes safe actions in uncertain situations (Martínez-Córcoles et al., 2012). Prior studies demonstrate the competencies required in Construction 5.0, such as digital literacy for managing safety analytics and cognitive flexibility for working alongside robots, complement traditional predictors like safety knowledge and motivation observed in manufacturing and healthcare sectors (Ohueri et al., 2024; Yu et al., 2021). Moreover, workers' trust in automation systems plays a crucial role in moderating safety participation (Strazzullo, 2024). However, the optimism associated with adopting new technologies in Construction 5.0 may inadvertently reduce safety participation if workers become overly reliant on automated safety measures (He et al., 2019).

In Construction 5.0, the fulfillment of safety-related job demands for workers relies on the combined support of emotional resources and technology-enhanced contextual resources. Drawing on the JD-R theory, safety participation can endow construction workers with contextual safety-related resources by cultivating a safety climate (Laurent et al., 2018). First, in collaborative human-robot work scenarios, the contextual resources, such as recognition mechanisms through real-time digital feedback systems, can assist workers in recognizing the safety behaviors that are encouraged by fostering positive safety perceptions (Laurent et al., 2018). This will eliminate the extraneous behaviors of construction workers and encourage them to concentrate more on safety engagement. Consequently, the waste of safety-related resources is reduced, which in turn generates additional opportunities to obtain additional safety practices. Second, psychosocial support, as a human-centric emotional resource, can boost the sense of team identification and job security among workers in Construction 5.0 Jiménez Rios et al., 2024; Laurent et al., 2018). This, in turn, motivates them to invest additional resources, such as those accessible through virtual reality (VR) training modules and real-time communication tools, in safety work, thus increasing their safety participation (Dollard et al., 2000; Guo et al., 2019). A high level of psychosocial support has been associated with enhanced safety participation (Guo et al., 2019; Higgins et al., 2010). Third, H1 posits that in Construction 5.0, psychosocial support can enhance safety practices by improving workers' utilization efficiency of psychosocial resources. Through safety participation, with the assistance of IoT and sensor technology, a psychosocial safety climate can be established. This climate encourages workers to engage in more safety activities, allowing for the dynamic allocation of safety resources according to their job responsibilities (Dollard et al., 2012). This demonstrates that safety participation in Construction 5.0 not only replenishes contextual resources through digital workflow integration but also ensures the efficient utilization of psychosocial resources for better safety practices. Based on the above-mentioned discussion, three hypotheses are proposed:

Leadership safety behavior refers to strategic organizational actions that foster employees' safety consciousness during task execution through structured managerial interventions (Cheng et al., 2020; Zhang et al., 2017). In the context of Construction 5.0, which integrates advanced digital technologies and automation, these actions include developing IoT-integrated safety protocols, implementing AI-driven risk assessment systems, conducting VR-enhanced safety training, and establishing real-time safety communication channels through digital collaboration platforms (Zhang et al., 2017). Within smart construction environments, leadership safety behavior not only demonstrates organizational competence in safety management but also reflects leaders' commitment to worker welfare through technology-mediated protective measures (Li et al., 2018; Tabassi et al., 2017). Leaders in Construction 5.0 face the challenge of managing human-robot collaboration and the deployment of IoT devices, both of which introduce new safety complexities (Ohueri et al., 2024; Zafar et al., 2024). Effective leadership safety behavior is essential for improving safety performance by ensuring that employees receive proper training in new technologies and that safety protocols are regularly updated to address emerging risks (Zhang et al., 2017). Additionally, this behavior fosters greater safety awareness among employees and encourages the correction of unsafe practices, particularly in the rapidly evolving environment of Construction 5.0 (Cheng et al., 2020). As a type of leadership safety behavior, leadership safety behavior is influenced by both internal and external factors (Zhang et al., 2017). Internally, leaders' risk preferences are shaped by individual safety attitudes, differences, motivations, and safety skills, which can influence their leadership styles and subsequently their leadership safety behavior (Wang and Zhao, 2019; Zhang et al., 2017). Externally, factors such as social capital, safety policies, and suppliers' requirements, which determine situational factors, can affect leadership safety behavior by shaping conditions outside the organizations (Courtright et al., 2014; Zhang et al., 2017). In Construction 5.0, these factors also include the integration of new technologies, regulatory compliance related to digital infrastructure, and the need for continuous adaptation to technological advancements (Ohueri et al., 2024; Shirwa et al., 2025).

As posited in H2, safety participation empowers construction workers to discern contextually appropriate safety behaviors within Construction 5.0 environments, effectively minimizing non-value-adding actions and resource misallocation while generating enhanced safety practice opportunities. Functioning as a techno-managerial interface, leadership safety behavior amplifies this mechanism through dynamic provisioning IoT-integrated safety protocols, AI-driven risk assessment systems, VR-enhanced safety training, real-time safety communication channels, and so forth (Hoffmeister et al., 2014; Tummers and Bakker, 2021). This integration not only solidifies an organization's commitment to safety but also helps workers develop a contextual understanding of safety objectives through immersive digital tools, thereby increasing the effectiveness of safety practices (McConville et al., 2016; Zhang et al., 2017). In line with the JD-R theory, safety participation (as psychological resource) and leadership safety behavior (as organizational resource) synergistically address the job demands of safety practice in Construction 5.0 (Xia et al., 2020). In this context, leadership safety behavior can be seen as an environmental factor that creates favorable conditions for safety participation and practices (Graham et al., 2020). Strong leadership safety behavior provides additional organizational resources that enhance safety participation by establishing various safety contexts through technology-driven interventions (Griffin and Hu, 2013; Xia et al., 2022a). This encourages workers to focus on safety activities and improves overall safety practices. Conversely, weak leadership safety behavior leads to fragmented digital safety networks and underutilized predictive maintenance systems (Griffin and Hu, 2013; Xia et al., 2018b). This lack of technological engagement reflects an organization's ambivalence toward safety, which can diminish workers' psychological commitment to safety activities and hinder their ability to meet the evolving safety demands of smart construction environments (Griffin and Hu, 2013; Neal and Griffin, 2006). Ultimately, leadership's ability to integrate digital and human safety resources critically determines the efficacy of safety participation-practice conversion. Thus, the fifth hypothesis is proposed as follows:

Figure 1 illustrates the conceptual model in accordance with H1, H2, H3, H4 and H5.

A questionnaire survey can not only collect massive data from target groups but also explore general views of construction workers on variables (Mohamed, 2002; Ng et al., 2005). Questionnaire surveys are widely applied to measure and analyze relationships between variables in construction projects (Hwang et al., 2018; Romm, 2013). This study aims to test the hypotheses specified in Figure 1. Therefore, a face-to-face questionnaire survey was conducted from April to August 2023 in order to test the hypothesized relationships between psychosocial support, safety participation, safety practice, and leadership safety behavior in Construction 5.0.

The questionnaires were initially prepared in English but were translated into Chinese to accommodate the workers' language needs, followed by back-translation into English for accuracy. The questionnaires consisted of two sections: the first focused on background information, while the second examined four key areas: psychosocial support, safety participation, safety practices, and leadership safety behavior. The survey concentrated on projects exemplifying the attributes of Construction 5.0, encompassing intelligent construction and smart construction sites. Before distributing the formal questionnaire, a pilot study was conducted in Nanjing to gather feedback from workers on the initial version. Their insights led to improvements in the clarity and readability of certain items, resulting in the finalized questionnaire. To ensure data reliability, a purposive sampling technique was employed (Ahmed et al., 2022). Additionally, considering that factors such as gender, experience, and education can influence psychosocial support, safety participation, safety practice, and leadership safety behavior in Construction 5.0 (O'brien et al., 2010; Vithanage et al., 2022), specific criteria were established for respondents: (1) they must be front-line workers in Construction 5.0, like intelligent or smart construction sites, (2) they should use intelligent tools like welding, plastering, and spraying robots and so forth, (3) both skilled and unskilled female workers were included, and (4) they should have prior experience with accidents. Consequently, locations meeting these criteria were selected to conduct the survey for this study on various Chinese construction projects.

Construction workers were able to complete the questionnaire in the project meeting room during their meal breaks, with significant support from the project managers. Participants were instructed to answer the questions based on their current or past project experiences. To encourage participation, each respondent received a small gift, consisting of a toothbrush and a towel, for completing the questionnaire. In total, 200 questionnaires were distributed, resulting in 118 valid responses for data analysis. The sample reflected a diverse demographic, encompassing various factors such as gender, age, occupation, work experience, education level, daily income, and project location (see Table 1).

Initially, Cronbach's alpha coefficients were calculated for safety practices, psychosocial support, safety participation, and leadership safety behavior measures using SPSS 23.0 to assess their reliability before testing the hypotheses. Next, confirmatory factor analyses (CFAs) were conducted with Amos 21.0 to evaluate the convergence and discriminant validity of the four constructs. Following this, regression models were developed to analyze the proposed relationships (H1, H2, and H3). Finally, the bootstrapping method, involving 5,000 resampling iterations, was used to assess the mediating and moderating effects outlined in hypotheses H4 and H5.

The Bootstrap method enhances the accuracy and power of mediation and moderation estimates by constructing confidence intervals (CIs) (Russell and Dean, 2000; Vance et al., 2015). This method establishes a 95% bias-corrected CI by resampling 5,000 times from the original sample, following the three-step process outlined by (Baron and Kenny 1986). Notably, this approach does not rely on the assumption of normally distributed compound coefficients (Edwards and Lambert, 2007; Vance et al., 2015; Xia et al., 2022a). A significant mediating effect is indicated when the CI does not include zero (Lau and Cheung, 2012). To test for moderating effects, three standardized predictors are required: the independent variable (e.g., safety participation), the moderator (e.g., leadership safety behavior), and their interaction (e.g., safety participation × leadership safety behavior) (Cohen et al., 2003; Preacher et al., 2007). A moderation effect is confirmed if the interaction term is statistically significant (Xia et al., 2022a).

Reliability testing is applied to examine the internal consistency of a questionnaire survey, and the degree of the internal consistency can be expressed by the values of Cronbach's alpha (Xia et al., 2022a). The value is considered adequate when the alpha value exceeds 0.70 (Zhang et al., 2020). The Cronbach's Alphas of the current sample are 0.895, 0.892, 0.917, and 0.913 for the safety practice, psychosocial support, safety participation, and leadership safety behavior, respectively.

First, series of CFAs were conducted in AMOS 21 to access the uncorrelated degree of the four constructs (e.g., safety practice, psychosocial support, safety participation, and leadership safety behavior). The results show that the four-construct model had a reasonable fit to the data [χ² = 212.66, p < 0.01; degree of freedom (Df ) = 102, comparative fit index (CFI) = 0.952, Tucker-Lewis index (TLI) = 0.928; standardized root mean square residual (SRMR) = 0.047; and root mean square error of approximation (RMSEA) = 0.096], and outperformed the following models: the two-construct model that collapses psychosocial support, safety participation, and leadership safety behavior (χ² = 670.66, p < 0.01; Df = 134, CFI = 0.768, TLI = 0.736; SRMR = 0.078; RMSEA = 0.185); three-construct model that collapses psychosocial support and safety participation (χ² = 647.13, p < 0.01; Df = 132, CFI = 0.778, TLI = 0.742; SRMR = 0.075; RMSEA = 0.183); three-construct model that collapses safety participation and leadership safety behavior (χ² = 632.49, p < 0.01; Df = 132, CFI = 0.784, TLI = 0.750; SRMR = 0.077; RMSEA = 0.180); three-construct model that collapses leadership safety behavior and psychosocial support (χ² = 644.348, p < 0.01; Df = 132, CFI = 0.779, TLI = 0.744; SRMR = 0.076; RMSEA = 0.182); and one-construct model that collapses all of the four construct (χ² = 723.822, p < 0.01; Df = 135, CFI = 0.746, TLI = 0.712; SRMR = 0.068; RMSEA = 0.193) (Hu and Bentler, 1999). The four-construct model performs better than other models, demonstrating its superiority in result analysis.

Second, averaging the variance extracted (AVE) illustrated their convergent validity by gauging the extent of their overlap. Convergent validity is established when the values of all the variables are greater than 0.50 (Garbarino and Johnson, 1999). The AVEs for safety practice, psychosocial support, safety participation, and leadership safety behavior are 0.743, 0.596, 0.653, and 0.752 respectively, indicating convergent validity.

G^* Power software was utilized to perform a post-hoc power analysis to assess the robustness of the study's findings. An effect size f² of 0.15, a significance level α of 0.05, and a total sample size of 118 are provided as inputs, yielding a critical F of 2.458 and a power of 0.928. Figure 3 displays the results of the post-hoc power analysis.

(Cohen 1988) posits that an effect size f² of 0.15 is indicative of a medium effect. Furthermore, a statistical power of 0.928 surpasses the commonly accepted minimum threshold of 0.8 for adequate power.

Our study provides novel insights into how psychosocial support enhances safety practices within the dynamic human-cyber collaboration framework of Construction 5.0. The statistically significant result (P < 0.001) supporting H1 reveals that psychosocial interventions specifically counteract the emotional resource depletion caused by intelligent system integration defining feature of Construction 5.0 environments. This demonstrates that replenishing workers' emotional reserves through multi-source social support (colleagues, relatives, natives) enables effective safety behavior implementation despite the cognitive demands of human-machine teaming and real-time data processing. The findings establish an emotional equilibrium mechanism that resolves the tension between safety assignment complexity (heightened by cyber-physical systems) and safety practice sustainability-addressing a critical gap in occupational health research within Construction 5.0 contexts. Unlike previous psychosocial studies focused on work-family balance or youth adaptation (Bates et al., 2019; Xia et al., 2018b), our age-diverse sample demonstrates how relational support networks mitigate technology-mediated stressors unique to Construction 5.0, while maintaining alignment with lean management principles through optimized safety resource utilization.

Our findings reveal that psychosocial support's efficacy in enhancing safety participation is fundamentally mediated through human-centric interfaces characteristic of Construction 5.0 ecosystems. The statistically robust relationship (p < 0.001) demonstrates that in environments featuring collaborative robotics and digital twin technologies, workers' engagement in safety practices relies on psychosocial mechanisms that address the challenges of human-technology collaboration. Specifically, this includes the need for greater job security in automated workflows and a stronger sense of team identity within hybrid human-robot teams. This evidence builds on the stress-buffering model proposed by (Wang et al. 2018) by integrating it with key elements of Construction 5.0: (1) real-time feedback for proactive interventions, (2) user-friendly interfaces for social support, and (3) a strengthened professional identity within dynamic teams. Crucially, our operationalization of Liu et al.'s (2020) psychological safety construct now incorporates the cyber-physical dimensions of Construction 5.0. Workers' sense of belonging is fostered by both human networks and their confident interactions with collaborative robots. Moreover, safety participation now encompasses not only physical precautions but also active involvement with predictive safety analytics platforms. This dual reinforcement mechanism, where intelligent systems enhance human psychosocial support and vice versa, establishes a new paradigm for safety management in cognitively demanding and technology-rich work environments.

The novel contribution of this study lies in revealing how safety participation operates as a catalytic converter within Construction 5.0's human-cyber-physical systems. Our findings demonstrate that H2 was statistically significant at the 0.001 level (P < 0.001), establishing that safety practice improvements in smart construction environments fundamentally depend on safety participation enhancement. This mechanism manifests through two distinct pathways aligned with Construction 5.0's analytics capabilities: (1) the optimized use of technology-driven safety resources, such as risk prediction and behavioral monitoring, and (2) improved identification and elimination of unsafe activities through collaborative workflows between humans and technology. These findings build on the work of (Laurent et al. 2018), which emphasizes the importance of organizational safety frameworks, by illustrating how worker participation actively engages human-cyber-physical safety systems. This engagement is crucial in Construction 5.0, where human decisions interact with autonomous equipment. Such operational clarity facilitates lean safety management by allowing for dynamic adjustments in behavior-practice trade-offs, particularly beneficial in projects that employ modular construction and sensor technologies. Furthermore, these findings offer computational validation for theoretical proposition of (Bayram 2020). They show that collaborative safety participation, as opposed to simple compliance, is a key driver of improvements in safety practices within technology-intensive work environments. This study clarifies previous uncertainties by providing empirical evidence from smart construction contexts, underscoring the critical role of active engagement in safety processes.

This study presents new evidence that, within the collaborative context of Construction 5.0, safety participation fully mediates the relationship between psychosocial support and safety practices (H4). This finding indicates that the influence of psychosocial support on safety behaviors is entirely channeled through workers' proactive engagement in safety activities. This engagement is crucial in environments characterized by real-time data sharing, advanced technology-driven risk prediction, and human-machine collaboration. Specifically, safety participation enables workers to effectively respond to the dynamic safety demands of Construction 5.0 by directing their emotional resources toward behaviors encouraged by a supportive safety climate. This discovery builds on previous research by (Kocatepe and Parlak 2021) and (Bunner et al. 2018), which examined accident prevention frameworks in smart construction ecosystems. It illustrates how safety participation shapes workers' understanding of safety protocols amid rapid technological changes. Additionally, it highlights how the role of safety participation expands to connect psychosocial factors with technological safety systems, a dimension that has been overlooked in business mediation models by (Yaris et al. 2020). Our empirical validation of this mediation pathway within the fluid organizational structures of Construction 5.0 provides a theoretical foundation for managing safety in temporary smart worksites, where human adaptability interacts with automated systems.

This study reveals novel empirical insights into how leadership safety behavior operates as a techno-managerial catalyst in Construction 5.0 environments. The findings confirm H5, demonstrating that leadership safety behavior positively moderates the relationship between safety participation and safety practice. In smart construction ecosystems, leadership safety behavior amplifies safety participation's efficacy through dynamic provisioning of AI-driven risk analytics, VR-enhanced training modules, and real-time safety communication via digital platforms. This technological mediation enables leaders to systematically reconfigure situational safety factors while embedding organizational safety values into workers' cognitive schemas through immersive digital interfaces. The results advance prior research by (Chu et al. 2021) and (Guo et al. 2019) on organizational safety support by delineating how Construction 5.0's digital infrastructure transforms leadership safety behavior into both a situational shaper (via IoT-integrated interventions) and a psychological anchor (through technology-mediated safety narratives). This dual mechanism strengthens the JD-R theoretical framework by showing how leadership safety behavior, as an organizational resource, interacts with safety participation to address emerging techno-physical job demands. Furthermore, the study extends spiritual leadership model proposed by (Liu et al. 2023), highlighting how digitally enhanced leadership safety behavior reshapes safety-related social norms within the construction workforce. It emphasizes that in Construction 5.0, leaders must effectively coordinate safety data, innovative training, and communication channels to prevent fragmented digital safety systems. This coordination is essential to ensure that workers' psychological safety aligns with evolving workplace demands.