One of the core characteristics of Industry 5.0 is the strong emphasis on human factor and human centricity, thereby prioritizing human well-being alongside technological advancements [1]. This paradigm shift aims to enhance human-machine interactions by creating more personalized, adaptable, and inclusive work environments [2]. Industry 5.0 focuses specifically on integrating human well-being and creativity with machine efficiency, fostering a collaborative environment where humans and robots can coexist harmoniously [3]. In this direction, the integration of collaborative robots (cobots) in industrial settings, particularly in manufacturing, is reshaping the landscape of modern production processes. Cobots are designed to work alongside human operators in a shared fenceless workspace and are meant to enhance productivity, safety, and efficiency in various tasks [4]. Cobots are deployed especially in assembly tasks, where the payload capacity and accuracy typical of traditional robotic systems need to be integrated with the skills and adaptability of human operators [5]. Indeed, unlike traditional industrial robots that operate in fenced environments, cobots are built to interact directly with humans. As such, they can be viewed as a dyad, engaging in both physical and cognitive interactions [6]. This emerging human-robot collaboration (HRC) aims to blend the strengths of both parties, creating a synergistic work environment that benefits from the unique capabilities of each [7]. Studies on social robotics have evidenced that the actual experience of human-robot interaction (HRI) can influence an individual’s attitudes, perspectives, and emotional responses regarding robots [8,9]. For instance, in social and educational contexts, HRI has been shown to foster positive attitudes toward robots and support individuals’ mental well-being [10]. However, in manufacturing settings, the deployment of robots, especially cobots, appears to have mixed effects on human operators. Recent studies highlight both the potential benefits and challenges associated with cobots [11]. On the one hand, cobots can relieve human operators from repetitive tasks, reducing physical workload and improving job satisfaction [12]. They are also found to enhance job performance by enabling operators to focus on more complex and engaging activities that require human creativity and problem-solving skills [13]. However, the presence of cobots can also introduce new demands and stressors, such as the need for constant monitoring, increased cognitive load due to task complexity, and potential job insecurity [13]. The aforementioned factors can negatively impact operators’ well-being if not properly managed [14].

As the deployment of cobots is becoming more widespread, it is essential to understand their impact on the operators’ affective well-being. According to the American Psychological Association [15], affective well-being encompasses emotional factors such as stress and depression and anxiety symptoms that may arise in operators exposed to advanced production technologies such as cobots. These factors are crucial in determining the success of the human-cobot dyad, referring to both the effectiveness of the working performance and the operators’ well-being. For example, anxiety symptoms and negative attitudes toward novel technologies can influence an individual’s trust in robots as coworkers [16]. This focus on affective well-being is particularly crucial in manufacturing, where the cognitive workload is known to be elevated [17]. Cognitive workload refers to the mental effort required during tasks, which, if excessive, can reduce job satisfaction, cause mental fatigue [18], and negatively impact operators’ affective well-being. Even though the definitions of HRI, and HRC are still blurred [19], it is essential to distinguish between the related terms often used in this field. HRI refers broadly to any form of interaction between humans and robots, encompassing both industrial and social contexts [8,9]. HRC is a subset of HRI that specifically involves cooperative task execution and goal sharing between humans and robots, often requiring a high degree of coordination [7]. Human-cobot interaction narrows the scope further to interactions with cobots, including both passive and active exchanges, without necessarily implying joint task execution. Human-cobot collaboration (HCC), which is the focus of this study, refers specifically to situations in which humans and cobots work together in a shared workspace on common tasks, combining physical or cognitive efforts [20,21]. In line with the study by Dhanda et al [20], in this review, we conceived the collaboration between humans and robots as reflected by their shared workspace and tasks. In addition, within this frame, the human-cobot dyad gains meaning not only as a task-driven entity but also as a social entity [21] that expresses itself both at an organizational and a relational level. The deployment of cobots is typically viewed primarily through the lens of productivity and safety enhancements [22]. The affective responses of workers interacting with advanced robotic systems such as cobots remain poorly understood and underreported. In this regard, previous reviews have mostly focused on the broader aspects of interaction and collaboration with industrial robots and not specifically on cobots, emphasizing cognitive and physical ergonomics over affective states. In particular, they have examined user experience [23], physical and cognitive ergonomics [24], mental workload [25], human factors [26], decision-making [27], and mental stress and safety awareness [28]. While several existing reviews have investigated HRI, ergonomics, and mental workload in industrial contexts, these works often either adopt a broad technological focus or limit their analysis to single dimensions such as physical ergonomics or user experience without specifically addressing the joint impact of cobots on both affective well-being and cognitive workload. Moreover, these reviews do not distinguish between traditional industrial robots and cobots, whose closer interaction with humans poses distinct psychological challenges. Unlike these prior works, this review uniquely synthesizes findings related to both cognitive and affective dimensions specifically within HCC, providing an integrative framework that highlights how cobots impact operators’ psychological experiences in manufacturing settings. Indeed, an overall analysis of the research addressing the effects of HCC on workers’ well-being is still missing.

Therefore, it seems relevant to conduct a systematic review of existing literature to better understand how working with cobots impacts operators’ affective well-being and how cognitive workload may contribute to these outcomes in manufacturing environments. By synthesizing the findings of the retrieved studies, this review intended to fill the identified gap in the literature, offering insights that could guide further research and the development of intervention strategies and best practices. These insights are relevant for ensuring that the technological benefits of cobots do not overshadow the organizational gains and the psychological factors of human operators, thereby aligning with the principles of Industry 5.0.

We conducted a comprehensive search of literature published up to April 2024 using 4 electronic databases—Web of Science, Scopus, ACM Digital Library, and IEEE Xplore—to identify relevant studies. Moreover, in the initial searches, we also adopted a manual snowballing method to detect further relevant publications by examining references from all systematic, literature, and scoping reviews. The search aimed to include all relevant peer-reviewed articles examining the cognitive workload and affective well-being of operators working with cobots in manufacturing environments. Search terms were identified and combined using Boolean operators to ensure thoroughness. The search terms included combinations of “cobot*,” “collaborative robot*,” “human-robot interaction,” “human-cobot collaboration,” “manufacturing,” “well-being,” “stress,” “anxiety,” “depression,” “cognitive workload,” and “mental workload” (refer to Multimedia Appendix 1 for the final search strategy string).

The inclusion criteria for selecting studies were defined as follows: (1) peer-reviewed articles published in English and for which the full text was available to the authors at the time of review (regardless of open access status); (2) studies focusing on the impact of cobots on operators’ cognitive workload in manufacturing environments; and (3) studies that provided empirical evidence on the effects of HCC on affective well-being, specifically anxiety, stress, and depression symptoms. The exclusion criteria were related to (1) studies focusing solely on technical aspects of cobots without considering the aforementioned dimensions; (2) studies that focused only on physical ergonomics related to the interaction with cobots; (3) studies not related to the manufacturing context; and (4) reviews, theoretical papers, and studies not directly examining the outlined keywords.

Eligibility screening was conducted in 2 stages: initially by assessing titles and abstracts, followed by full-text review of potentially relevant studies. Two independent reviewers (GB and Giuseppe Donà) carried out the selection process. Any discrepancies were discussed and resolved by including a third reviewer (SS).

Data extraction was conducted independently by 2 reviewers (GB and GD). Any discrepancies in the extracted data were discussed and resolved through consensus, with the involvement of a third reviewer (SS) when necessary.

The extracted data included (1) study characteristics (title; author name; year of publication; publication type [journal or conference proceeding]; keywords; country; study objectives; study design; and sample size, mean age and SD, and gender identity); (2) the type of advanced robotic systems; (3) the type of tasks performed by the human operator and cobot; (4) the type of simulating scenarios used; (5) measures for evaluating affective well-being, in particular anxiety, stress, and depression symptoms, as well as cognitive workload; and (6) key findings and limitations of the included studies related to the purpose of this systematic review.

Due to the expected heterogeneity in study designs and outcomes, a narrative synthesis was conducted, allowing for the integration of qualitative and quantitative data. We identified key factors that impact operators’ affective well-being and cognitive workload, placing them in the broader context of HCC in manufacturing.

This study received approval from the ethics committee of the Human Inspired Technologies Research Center of the University of Padua, Italy (2023_212R1).

This systematic review adhered to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [29] to ensure a rigorous and transparent review process (Figure 1 [29]). The initial search yielded 2393 records. After removing duplicates (n=146, 6.1%) and records excluded for other reasons (n=2194, 91.7%), 53 records (2.21%) remained for screening. Following title and abstract screening, 3 records (5.7%) were excluded, leaving 50 reports for retrieval. Of these, 3 reports could not be retrieved, resulting in 47 reports assessed for eligibility. On the basis of the inclusion and exclusion criteria, 33 reports (70.2%) were excluded, leaving 14 studies for inclusion. Using the snowballing method, an additional 32 articles were deemed eligible, resulting in a final total of 46 articles included in the systematic review.

We systematically reviewed 46 studies published between 2009 and 2024, revealing significant trends and insights into the impact of cobots on human operators’ affective well-being and cognitive workload. Notably, there was a marked and constant increase in publications from 2020 onward, with a peak in 2023, likely reflecting the growing interest in understanding the human factors in HCC. Most studies (25/46, 54%) were conducted in Italy and the United States, primarily in laboratory settings and without involving real shop floor workers. Indeed, participants were mostly male university students or laboratory members aged approximately 29 years. This prevalence of controlled environments and nonrepresentative samples significantly limits the generalizability of the findings. Laboratory studies offer useful baseline insights, but they fail to capture the complexity, unpredictability, and organizational dynamics of real-world industrial settings. In actual workplaces, factors such as shift patterns, production pressure, diverse skill levels, and sociotechnical constraints may interact with cobot integration in ways that are not observable in laboratory conditions. Moreover, real-world HCC is embedded within social structures in which workers must navigate interpersonal relationships with managers and colleagues [21]. Cobot adoption may influence these dynamics, affecting perceptions of fairness, trust, role clarity, and social cohesion on the shop floor. For example, workers might experience increased stress not only from the cobot itself but from unclear expectations, fears of job replacement, or tensions in team collaboration [21].

Moreover, the most used advanced robotic system was Universal Robots through different weight categories used particularly for assembly and disassembly tasks, thereby failing to reflect the variety of robotic systems actually deployed in real-world settings.

In particular, the included studies primarily focused on exploring aspects such as cognitive or mental workload, human factors, and occupational safety using physiological measurements and self-report tools either alone or combined. Researchers have paid considerably less attention to emotions, affective states, well-being, and symptoms of anxiety and stress. It is worth noting that none of the included studies examined depression symptoms, which are critical indicators of chronic stress [78-80]. This imbalance highlights the necessity for future research to explore the emotional and psychological aspects of working with cobots more thoroughly. By understanding these dimensions, companies can develop strategies that enhance the affective well-being of human operators while simultaneously maintaining or improving productivity and efficiency. This, in turn, can inform the design and implementation of cobot systems that not only support operational goals but also foster a healthier, more supportive work environment for operators.

In terms of affective well-being, the studies observed that, while cobots have been shown to alleviate physical fatigue and improve job satisfaction, they also introduce new psychological challenges. These include stress and anxiety symptoms stemming from concerns about job security and the stress that comes with high-paced operations. Critical factors influencing these outcomes include the modality of the introduction of cobots within the workplace, the proximity of cobots, the speed at which cobots operate, the system usability, and the complexity of tasks assigned. In particular, Brun and Wioland [30] found that operators experienced fewer stress symptoms and physical discomfort when cobots were newly introduced. Moreover, operators’ stress levels increased significantly when cobots operated at higher speeds or in close proximity [44-46,56,57,61,62,69]. In this regard, findings may suggest that limiting the manipulator speed to lower levels, preferably <500 mm/s, could mitigate stress symptoms and enhance operator comfort. Indeed, providing operators with the ability to control cobot behavior, such as adjusting speeds, can mitigate these stress responses and enhance overall comfort [57]. In addition, effective system usability significantly impacts operator stress levels, particularly through communication interfaces between cobots and human operators. Arntz et al [47] showed that augmented interfaces (eg, text panels, light signals, and gestures) decreased perceived stress and frustration, whereas Amanhound et al [32] highlighted that clear communication and timely feedback from cobots further alleviated stress and enhanced task performance. In this regard, it is also important to consider the task complexity, which significantly impacts stress [53,55] and anxiety levels. For instance, Rajavenkatanarayanan et al [42] highlighted that the complexity introduced by task timers and continuous work heightened stress symptoms. Eimontaite et al [52] noted that task accuracy and signage reduced anxiety symptoms.

In terms of cognitive workload, studies found similar results, pointing out that higher task complexity increased cognitive load, with dual-task conditions significantly heightening mental workload compared to single-task conditions [36,38,49,55,57,59-61,64,71]. Moreover, studies indicated that user control over cobot behavior, such as adjusting speed, alleviated mental fatigue [54,75], whereas high-speed conditions increased mental workload [43,46]. Furthermore, collaborating with cobots generally reduced both mental and physical demands compared to manual workstations [62], although the cobots’ effect on participants’ cognitive workload remained mixed [70]. Collaboration with cobots did not generate greater mental load compared to human collaboration [72]. However, it can make users feel alienated despite increased focus time [31]. In this regard, effective communication interfaces and adaptive systems that adjust cobot actions based on real-time assessments of human stress levels and cognitive load are promising strategies to improve HCC quality and, thus, their fluency, meaning the smoothness and quality of their interaction. Indeed, studies highlighted that integrating cognitive aids [74], usability guidelines [59], affective communication [37], trust-based frameworks [39], interaction modalities [48,65,66], and motion-based signals [33] all contributed to lower cognitive workload. Another important aspect that emerged from this systematic review concerns the considerable heterogeneity in how affective well-being was measured across the studies. Tools ranged from validated self-report scales to physiological proxies such as HRV. The lack of consensus on which instruments best capture affective responses to HCC hinders replication and slows cumulative progress in the field. To address this issue, future research should focus on the development and adoption of a standardized, multidimensional battery for assessing affective well-being in HCC contexts. This could include widely used and validated psychometric tools (eg, the 7-item Generalized Anxiety Disorder scale for anxiety, the 9-item Patient Health Questionnaire for depression symptoms, and the PSS for perceived stress) supplemented by context-sensitive physiological measures and qualitative interviews to triangulate findings and ensure robustness.

As this is the first systematic review specifically focused on the affective well-being and cognitive workload of operators in HCC, it was crucial to first establish a comprehensive narrative mapping of the existing evidence. This foundational step aligns with the best practices recommended for early-stage systematic reviews in nascent research domains [76]. While a meta-analysis can provide statistical rigor and pooled effect estimates, it was not feasible to conduct one in the context of this review. The motivation was primarily the high degree of heterogeneity across the included studies, particularly in terms of study designs, participant populations, settings (eg, controlled laboratories vs industrial environments), measures used (eg, NASA-TLX, HRV, EEG, and various self-report scales), and outcomes reported. In addition, many studies reported incomplete or nonstandardized statistical data, such as missing effect sizes or inconsistent outcome reporting, which further limited the feasibility of a quantitative synthesis. This limitation is consistent with what has been noted in other systematic reviews in emerging interdisciplinary fields, where methodological variability and limited high-quality data preclude meaningful meta-analytical aggregation [77,81]. This systematic review highlights that cobots offer substantial benefits in terms of productivity and safety in manufacturing but also that they undoubtedly weigh on the workers’ mental workload, especially when the situational complexity is high. However, the fact that nearly all studies (44/46, 96%) were set in controlled laboratory settings and with the involvement of nonrepresentative participants and tasks highlights the need to conduct experiments in more ecological conditions (ie, in industrial settings). This would help research better reflect the diverse characteristics and competences of workers and the full range of activities they perform. Real-world scenarios will pose significant challenges to the robustness of physiological measurements, which were frequently used in the reviewed studies. To overcome this, future research may integrate other research methods (eg, video analysis, as done in the study by Kopp et al [82] to assess mental workload in a manual assembly task). Future research should also broaden its scope to include the evaluation of depressive symptoms alongside anxiety and stress symptoms as all are critical indicators of operators’ affective well-being. Addressing these psychological factors using similar standardized measures can significantly enhance overall affective well-being, leading to increased productivity in the organization. Moreover, the addition of a qualitative approach (eg, interviews) would allow for better understanding of the workers’ experience and would also compensate for the substantial lack of qualitative methods in the retrieved studies. Studying cobots in real-world settings would also make it possible to evaluate those with a high technology readiness level, which are technological solutions that can actually be deployed in manufacturing.

Consistently with the paradigm of Industry 5.0, these systems should be designed placing the human operator at the forefront. To this end, workers should be enabled to control the features of the cobots’ movements whenever possible. In addition, we recommend that manufacturing companies integrate cobots gradually involving workers throughout the process and using pilot phases that include both technical training and psychological monitoring of workers, as already highlighted in the study by Kopp et al [82], given the proven impact that interacting with cobots can have on workers’ cognitive and affective well-being. Human resources departments are encouraged to include mental health assessments in routine occupational evaluations to monitor the fitness of the HCC and, eventually, be able to promptly address emerging issues. In addition, interdisciplinary collaboration between engineers, psychologists, and safety officers should be institutionalized to co-design adaptive cobot systems that account for the best interests of both organizations and workers. These strategies will help transform technological innovation into sustainable workplace practices that align with the human-centered goals of Industry 5.0. Finally, research efforts should be focused on developing systems capable of dynamically adjusting cobot behaviors in response to human operators’ stress and anxiety symptoms as well as cognitive workload. This approach will support the physical aspects of HCC while also prioritizing the operators’ affective well-being.

This systematic review highlighted key indicators related to the impact of cobots on the affective well-being and cognitive workload of human operators. The analysis of these studies revealed that, while cobots alleviate physical fatigue and improve job satisfaction, they also introduce new psychological challenges, such as increased stress and anxiety symptoms, particularly during high-speed and proximity operations. In addition, increased task complexity and dual-task conditions significantly increase cognitive workload. Furthermore, the results of these studies emphasized the potential of adaptive systems, which allow for the real-time regulation of cobots’ behavior and provide operators with more control as these systems can significantly reduce stress symptoms and cognitive workload, thereby improving overall affective well-being and efficiency. Notwithstanding these promising results, the generalizability of these findings is limited by small, homogeneous samples and controlled laboratory environments. Therefore, future research should address these limitations by diversifying participant samples and conducting studies in real industrial environments to improve ecological validity.

In addition, the consistent use of the NASA-TLX to assess cognitive workload in all these studies ensured comparability of results, demonstrating a clear trend of increasing cognitive demands with more complex tasks. However, the assessment of affective well-being was less homogeneous, with the adoption of a wide variety of measures, which makes it difficult to compare results across studies and underlines the need for more consistent assessment tools in this domain.

Therefore, it is crucial to standardize the use of these measures as it will potentially improve methodological rigor and facilitate comparative analyses between studies. In this regard, the combined use of physiological measures and self-report tools to comprehensively assess stress, anxiety, and depressive symptoms along with cognitive and physical workload assessments is crucial. Furthermore, there is the need to integrate qualitative methodologies that provide insights into the lived experiences of operators working with cobots. By capturing the subjective experiences of these operators, the broader implications of the use of cobots in manufacturing can be better understood.

Mindful of this, a holistic approach to well-being is essential. This approach recognizes that well-being includes not only the alleviation of symptoms of stress and anxiety but also the management and reduction of depressive symptoms and other crucial psychological factors. For example, it is necessary to comprehend the individual as a complex human being, including the emotional, mental, and social dimensions. By addressing these aspects, it is possible to develop a more nuanced understanding of the psychological impact of cobots on human operators and subsequently implement strategies to more effectively support the affective well-being and cognitive workload of operators working with cobots.

Future research should aim to develop longitudinal studies that track psychological responses to cobots over time, ideally in real industrial environments rather than controlled laboratory settings. Samples should include diverse worker populations such as shift workers, older adults, and nontechnical staff to improve ecological validity and generalizability. Mixed methods approaches that combine quantitative measures (eg, HRV, EEG, and NASA-TLX) with qualitative methods (eg, interviews) will be critical to fully capture the complexity of HCC. In addition, experimental designs that manipulate variables such as cobot autonomy, task complexity, or interface transparency can help isolate causal effects on cognitive workload and affective well-being.

Therefore, the development of these strategies will create healthier and more supportive work environments and refine the design and implementation of cobot systems, thus ensuring greater efficiency and productivity.