Abstract
Purpose
Most studies have focused on the impact of the application of AI on management attributes, management decisions and management ethics. However, how job demand and job control in the context of AI collaboration determine employees' learning process and learning behaviors, as well as how AI collaboration moderates employees' learning process and learning behaviors, remains unknown. To answer these questions, the authors adopted a Job Demand-Control (JDC) model to explore the influencing factors of employee's individual learning behavior.
Design/methodology/approach
This study used questionnaire survey in organizations using AI to collect data. Partial least squares (PLS) predict algorithm and SPSS were used to test the hypotheses.
Findings
Job demand and job control positively influence self-efficacy, self-efficacy positively influences learning goal orientation and learning goal orientation positively influences learning behavior. Learning goal orientation plays a mediating role between self-efficacy and learning behavior. Meanwhile, collaboration with AI positively moderates the impact of employees' job demand on self-efficacy and the impact of self-efficacy on learning behavior.
Originality/value
This study introduces self-efficacy as the outcome of JDC model, demonstrates the mediating role of learning goal orientation and introduces collaborative factors related to artificial intelligence. This study further enriches the theoretical system of human–AI interaction and expands the content of organizational learning theory.
Keywords
Citation
Chen, A., Yang, T., Ma, J. and Lu, Y. (2023), "Employees' learning behavior in the context of AI collaboration: a perspective on the job demand-control model", Industrial Management & Data Systems, Vol. 123 No. 8, pp. 2169-2193. https://doi.org/10.1108/IMDS-04-2022-0221
Publisher
:Emerald Publishing Limited
Copyright © 2023, Emerald Publishing Limited
1. Introduction
The efficiency of an organization depends on whether everyone in the organization can make the best use of their own talents and knowledge. Employee learning behavior plays a crucial role in organizational innovation and organizational change. At the same time, the application of artificial intelligence (AI) in organizations is one of the most influential drivers of organizational innovation and organizational change. More and more enterprises have begun to use AI as a tool to improve work efficiency. Some organizations have even adopted AI as human-autonomy teams (HATs). Some scholars refer to the AI era as the “Industry 4.0 Revolution” and believe that AI will be a crucial element in future work. AI enhances human capabilities through learning, perception, and cognition, reducing errors and increasing efficiency, speed, and accuracy for organizations. This highlights the interaction between machines and people, with AI becoming a key driver of business model innovation and process change, leading to enhanced competitiveness (Sima et al., 2020). The study focuses on exploring the impact of human–computer collaboration by examining the development of employee learning behavior in the context of AI collaboration. The core goals of AI are twofold: to understand human intelligence and thinking, and to be able to process complex information tasks with human or even beyond human intelligence (van der Maas et al., 2021). Complex processes can be simplified in collaboration with AI assistants, which change employees' workflows and learning environments. With the help of AI, employees can complete work tasks more easily by using office automation tools. And companies are increasingly focused on how to automate routine tasks more effectively. In this case, will employees realize the importance of learning, innovation, and other skills, and thus have more willingness and more opportunities to learn more skills on the job? Or, given the continuous application of brain–computer interfaces, will employees give the human brain a complete vacation? The utilization of AI technology in organizations to shape employee learning behavior is prevalent. For instance, some companies employ AI to devise tailored training programs and provide intelligent coaching and feedback mechanisms to enhance employees' learning effectiveness and enthusiasm (Kuleto et al., 2021). Such an approach can lead to enhanced job performance and increased productivity, benefiting both the individual employee and the organization as a whole. Additionally, AI can also enhance employees' learning experiences and engagement by utilizing adaptive learning systems, gamified learning and virtual reality (Alam, 2022). However, when implementing AI, employees and organizations may encounter challenges. For example, employees may face difficulties in mastering new technologies or lack trust, which may suppress their willingness to learn (Schilling and Kluge, 2009). Moreover, since the application of AI often requires a higher level of technical skills and professional knowledge, they may increase employees' learning burden and induce psychological pressure. In this case, whether AI promotes or suppresses learning engagement depends on specific situations and application contexts. If AI applications can effectively address work-related issues and offer engaging and valuable learning opportunities, then employees are more inclined to accept and participate in learning activities. However, if AI applications are overly complex or do not meet employees' expectations and needs, they may impede their willingness to learn. Regarding the continuous application of brain–computer interfaces, further research is required to explore its impact on employee learning. Some studies indicate that brain–computer interface technology can promote learning by improving attention and cognitive load, but it may also increase employees' pressure and fatigue, thus diminishing their motivation to learn (Gurkok and Nijholt, 2012). Therefore, it is necessary to consider employees' psychological and physical health when implementing brain–computer interface technology, to ensure the positive effects of learning outcomes and experiences.
At the same time, we face significant gaps in our knowledge of employees' learning behavior in the context of AI collaboration. In the field of employee learning, researchers have emphasized the importance of individual employee learning behavior and its benefits to the organization (Edmondson et al., 2007; Perry-Smith, 2006). In 1979, Robert Karasek proposed the Job Demand Control (JDC) model, which outlines the impact of job characteristics on stress, health and occupational well-being. Karasek regards job demand and job control are crucial workplace characteristics that affect employee welfare, motivation and productivity, and various physical and psychological pressures. The JDC Model is one of the most influential models for human resource-related job stress, organizational behavior and job design issues. It is widely used to assess the work-related stress and an individual's ability to cope with it by classifying four types of work stressors into high pressure, low pressure, passive and active work based on the combination of work demand and control level. The model suggests that the degree of difficulty and the amount of freedom employees have to execute behaviors at work jointly determine the stress levels of them. Since its inception, many scholars have questioned this theory and continuously developed and improved it (Häusser et al., 2010; Johnson and Hall, 1988). And the JDC model was proposed and examined to explain the mechanism of employee learning behavior in various contexts (Doef and Maes, 1999). JDC model seems to be related to the study of learning behavior in the field of employee learning, but it has rarely been studied in previous studies. Therefore, this paper will use JDC model as the overall theoretical perspective. However, the process that the job demand and job control formulate employee learning behavior has not been identified. The JDC model has long been considered a stable and widely accepted framework. The “learning theory” contained in it has explained the personal learning tendency of employees and the learning behavior formed by the tendency to some extent. Self-efficacy refers to an individual's ability to judge, believe or grasp and feel whether he can complete a certain task at a certain level (Graham and Bray, 2015). Self-efficacy does not refer to one's actual work abilities, but reflects the degree of confidence in utilizing them to complete tasks. Learning goal orientation refers to individuals' tendency to learn new skills and adapt to new situations. This study posits that learning behavior is determined not only related to the interaction between job requirements and control, but also individual self-efficacy and learning goal orientation, which affect how well job control matches job requirements. Therefore, we suggest that the traditional JDC model should incorporate individual self-efficacy and learning goal orientation. Inspired by the research on the employees' self-efficacy (Yoon and Kayes, 2016) and learning goal orientation (Chughtai and Buckley, 2010) playing important roles in determining, we propose integrating these factors into the JDC model.
On the other hand, the work environment also determines the formulation of employee learning behavior, while the burgeoning AI applications create an emerging work environment named AI collaboration. In the field of AI management, existing studies have focused on the impact of the application of AI on management attributes, management decisions, and management ethics (Kevin and Ana, 2019; Shrestha et al., 2019; Sowa et al., 2021). Despite the increasing integration of AI into enterprises and its assistance in completing work-related tasks, few studies have examined the impact of AI collaboration on the development of employee learning behavior. Therefore, it is essential to investigate the mechanism of employee learning behavior in the context of AI.
To fill these gaps, we adopted an extended JDC model to explore the influencing factors of employee's individual learning behavior in the context of AI collaboration. Using data from a large sample survey, we found positive relationships between job demand, job control, self-efficacy, employee learning behaviors and the mediating role of learning goal orientation. Collaboration with AI positively moderates the positive impact of employee self-efficacy on learning behavior, and job demand on self-efficacy. Here, we introduce self-efficacy as the outcome of JDC model, demonstrate the mediating role of learning goal orientation, and put forward some guidelines for organizations who want to enhance the learning behavior of employees in the context of AI collaboration.
2. Literature review and theoretical foundation
2.1 Literature review
In this section, we review the literature on employee learning behavior and the application of AI collaboration in organizations.
A large number of studies in the field of employee learning behavior focus on individual and team learning behavior. Edmondson (1999) proposed that team psychological security is related to learning behavior. Researchers then began to explore the generation of employee learning behavior at the psychological level, highlighting voluntary learning behavior. Personal self-efficacy based on organizational self-esteem and role breadth is positively correlated with individual voluntary learning behavior in the team (Akram et al., 2017; Kim and Lee, 2015; Yoon and Kayes, 2016). Therefore, we believe that self-efficacy is one of the factors affecting the individual learning behavior of employees.
Chughtai and Buckley (2010) and Hirst et al. (2009) discovered that learning goal orientation plays a mediating role, and identified a non-linear interaction between individual learning orientation and team learning behavior. Specifically, in teams with higher levels of team learning behavior, the positive relationship between learning orientation and creativity was found to be weakened at higher levels of learning orientation. Therefore, based on the learning hypothesis in the JDC model and referring to Salanova et al. (2002), the current study took self-efficacy and learning goal-orientation as intermediate variables between the JDC model and employees' learning behavior.
However, few studies have focused on the influence of organizational work environments on employee learning behavior. The application of AI is a representative example of the change in organizational work environment and learning environments. More and more researchers have begun to pay attention to the application of artificial intelligence in organizations, including the impact of the application of AI on management attributes, management decisions, and management ethics. For example, AI is fundamentally changing organizational decision-making processes (Cao et al., 2021b; Vincent, 2021), and the future of work is human-AI symbiosis (Jarrahi, 2018). (Chowdhury et al., 2022; Mikalef and Gupta, 2021) have examined the relationship between AI-employees collaborative ability and organizational creativity and business performance. The research on AI collaboration has mainly focused on how AI performs identity prediction (Mirbabaie et al., 2022), its impact on B2B sales channels (Paschen et al., 2020), and the relationship between AI and human resource management (Ahmad et al., 2021). These studies have significantly contributed to the development of contemporary team and AI collaboration. However, research on the influence of AI collaboration on employee learning processes and learning behavior is rare. Employees' learning behavior is a core element of organizational development, as it is closely related to the organization's overall growth and progress. Therefore, we take AI collaboration level as a moderating variable to explore its moderating effect on learning processes and learning behavior in organizations.
2.2 Presage–process–product (3P) model
Biggs (1993a) proposed the 3P model (adapted from Dunkin and Biddle (1974), which is a key framework in Student Learning Theory. By specifying relations between presage, process and product, the 3P model explains the relevance of students' learning methods from the perspective of dynamic systems (Biggs, 2001).
Presage factors are generally divided into two components: students' individual learning characteristics and learning environment (e.g. student prior knowledge, abilities, values, and expectations concerning achievement). Process variables include learning motivations and/or strategies. Product factors are the subsequent learning outcomes (Biggs, 1993b; Fryer and Ginns, 2018).
The 3P model stipulates that, in addition to the associations between presage factors and product factors, there are indirect effects generated from Presage variables, through Process variables, to Product variables. Process variables play an intermediary role in it. These indirect effects are held to be generated to the extent that meta-learning is activated (Baard et al., 2004; Biggs, 1993a).
Within this body of theory and research, Ginns et al. (2014) took students' self-efficacy and teacher support as presage factors and learning strategies as process factors, and verified learning strategies play a mediating role between self-efficacy and learning product. Students' self-efficacy is related to values and expectations concerning achievement, while their learning motivation and strategy depend on learning goal orientation to a certain extent. Based on these, we applied the 3P model to explore the learning behavior of employees and propose self-efficacy – learning goal orientation – learning behavior model.
2.3 Job demand-control model (JDC)
The JDC model, which was put forward in 1979, has been used in work and management studies in recent years. First, job demand refers to the requirements of the working situation and is generally measured in terms of time pressure and role conflict (Doef and Maes, 1999). Time pressure refers to the fact that employees to face too many tasks in too little time at work, while role conflict refers to the fact that employees are confronted with conflicting requirements from all sides at work. Second, job control refers to the ability of individuals to control work and tasks (Doef and Maes, 1999). It is generally believed that job control can be measured from two dimensions: the diversity of individual work skills and the decision-making power of the individual over their work content.
On this basis, the JDC model divides both job demand and job control into high and low categories according to the perception level. Different matching categories will have different impacts on employees' behaviors. High-high matching conditions can cause employees to have positive work experiences, low-low matching conditions will lead to negative work experiences, low-high matching situations will produce low working pressure and high-low matching conditions will lead to high work pressure. Karasek (1979) demonstrated the model in graphic form, as shown in Figure 1.
Based on this model, Karasek (1979) proposed two mechanisms that are represented by the diagonals. First, following the “strain” diagonal, strain increases as job demands increase, relative to decreasing job decision latitude. Second, following the “learning” diagonal, high demands, in combination with high control, lead to increased learning, motivation and development of skills (Doef and Maes, 1999).
Inspired by the second mechanism, this study explored the relationship between JDC and employee learning behavior, and considered the employees' beliefs and needs (i.e. self-efficacy and learning goal-orientation) as the mediators between JDC and employee learning behavior.
2.4 Self-efficacy
Self-efficacy is defined as “beliefs in one's capabilities to mobilize the motivation cognitive resources, and courses of action needed to meet given situational demands” (Bandura and Wood, 1989). Bandura (1986) argued that human behavior is not only influenced by the results of behavior, but also by the expectation of self-behavior ability and behavioral results formed through human cognition. Although it does not directly reflect the work ability of employees, the expected role of self-efficacy will change the judgment and choice of employees, and therefore affect their behavior. Current research on the influence of self-efficacy, Bandura (1977) believed that self-efficacy would affect individuals' choices, persistence time, and thinking mode. Specifically: (1) Self-efficacy will affect people's behavioral choices. People with a high level of self-efficacy believe that they can complete high-level and difficult tasks and work, and take this as a demonstration of their self-ability. (2) Self-efficacy will also affect people's decisions to stick with or give up on something difficult. (3) Self-efficacy has a great influence on individuals' ways of thinking. Newman et al. (2018) studied and verified the effects of employees' self-efficacy on innovative behavior.
Taking the above effects into consideration, we took self-efficacy as the mediation variable between the JDC model and learning behavior.
3. Research model
3.1 Job demand-control and self-efficacy
First, in discussing the theory of basic psychological needs, we mentioned that an employee's competence requires that they hope to succeed in the most appropriate challenging task and achieve the desired results. Deci and Ryan (1985) also pointed out that high-level job control can provide employees with abundant executable operations and coordinately allocated resources, so that employees can develop a corresponding sense of self-identity. Such growth, as a successful experience, can promote the generation of self-efficacy of employees and produce positive regulatory feedback.
Second, Dweck and Reppucci (1973) found that the increase in job demand may cause psychological pressure on employees, but also encourages employees to work hard to complete their tasks. Dweck (1986) argued that when holding a job performance goal, people would strive to improve their mastery and ability, and the process of mastering knowledge and the result of completing the performance goal could prove their ability to others, thus giving people a sense of self-esteem.
Therefore, we proposed the following hypotheses:
The level of job demand will positively affect the self-efficacy of employees.
The level of job control will positively affect the self-efficacy of employees.
3.2 Self-efficacy, learning goal orientation and learning behavior
As for the influence of employees' learning behavior, this study began from the assumption that employees' self-efficacy will affect their cognition of the importance of learning tasks at work—that is, their level of learning goal orientation.
On the one hand, people with high levels of self-efficacy believe they are able to complete high-level, difficult tasks, and are more likely to think of successful outcomes (Bandura, 1977). This allows them to have a higher learning goal orientation. On the other hand, they are often willing to work hard to overcome difficulties to complete the target task, and it is easier to turn ideas into reality and form learning behaviors. In contrast, people with low self-efficacy are more likely to think about the results of failure when facing difficulties, and it is easier for them to choose to give up rather than stick to a task.
Therefore, we proposed the following hypotheses:
Employees' self-efficacy will positively affect their learning goal orientation.
Employees' self-efficacy will positively affect their learning behavior.
The learning goal orientation of employees is influenced by their supervisors' assumptions about the expected results of learning, which plays a guiding role in their learning behavior. Higher learning goals can motivate employees to put in more efforts and produce better learning results.
Therefore, we proposed the following hypothesis:
Employees' learning goal orientation will positively affect their learning behavior.
In previous studies, Biggs (2001) found that differences in cognitive style, personality, values and other factors would lead to different learning styles and then determine learning outcomes. In this process, learning style is the process variable. Inspired by the above, and based on the 3P model, we took employees' learning goal orientation as the Process variable. We hypothesized that in addition to the direct positive relationship between self-efficacy and learning behavior, there is also an indirect positive relation from self-efficacy, through learning goal orientation, to learning behavior.
Therefore, we proposed the following hypothesis:
Employees' learning goal orientation plays a mediation role in the relationship between their self-efficacy and learning behavior.
3.3 The moderating effect of AI collaboration
Our model was developed according to the 3P model to elaborate the formulation process of employee learning behavior. The AI collaboration creates a specific working context in which employees formulate their learning behavior. This context factor can only change the formulation process of employee learning behavior. Previous studies have revealed the moderation roles of other similar context factors, including social media context (Cao et al., 2021a) and culture context (Diallo et al., 2018).
Research has shown that the application of AI can evoke job insecurity in employees (Koo et al., 2021), but the correlation decreases with the improvement of employees' career development level (Nam, 2019). Therefore, we can further assume that employees tend to have more positive work behaviors under high demand and high control. The collaborative application of AI can properly simplify the work tasks of employees, relieve the pressure caused by high requirements, and increase the confidence of employees in completing difficult tasks. When AI collaboration is high, pressure from high job demand is reduced, and employees are more motivated and have a higher sense of self-efficacy. At the same time, when the work tasks of employees are simplified, employees are more confident to complete difficult tasks, and their sense of self-efficacy is improved. Therefore, we propose the following hypotheses:
Collaboration with AI positively moderates the positive impact of job demand on employees' self-efficacy.
Collaboration with AI positively moderates the positive impact of job control on employees' self-efficacy.
The arrival of the era of artificial intelligence and the application of Internet big data can provide employees with more and richer paths and methods to acquire knowledge. At the same level of self-efficacy, more and richer learning approaches and methods may produce better employee learning behavior. In other words, the impact of employee self-efficacy on learning behavior will be influenced by AI collaboration, and the influence will be positive.
Therefore, we put forward the following hypothesis:
Collaboration with AI will positively moderate the positive impact of employees' self-efficacy on learning behavior.
3.4 Job demand-control, self-efficacy and learning goal orientation
Bandura discussed how beliefs about one's own abilities could affect motivation, cognitive processes, and behavior (Bandura, 1994). In uncertain environments, individuals with a low sense of control over the environment are less likely to gain motivation, creativity, and perseverance when facing obstacles, while those with a high level of self-efficacy are more likely to overcome challenges and pursue learning goals (Peng and Mao, 2015). This study proposes the relationship between job matching and job satisfaction, in which self-efficacy plays a mediating role. Research has found that self-efficacy can partially adjust the impact of person job matching on job satisfaction (Bancans et al., n.d). The impact of personal characteristics (self-efficacy, learning goal orientation, and extroversion) on learning activity and job satisfaction was examined. The author suggests that individuals with high self-efficacy are more likely to pursue learning goals. Therefore, based on the 3P model, it is hypothesized that self-efficacy beliefs can mediate between environmental factors such as job demands and control and personal outcomes like learning goal orientation.
Therefore, we proposed the following hypothesis:
Employees' self-efficacy plays a mediation role in the relationship between their job demand-control and learning goal orientation.
All hypothetical relationships for H1-H10 are shown in Figure 2.
4. Methodology
4.1 Measures
This study used questionnaire survey in organizations using AI to collect data. In designing the questionnaire, we adopted the relatively authoritative and reliable maturity scale to measure variables in the model, so as to ensure high reliability and validity and reduce errors. The questionnaire consisted of six parts. The specific scales of the questionnaire are shown in the Appendix Measurement Items.
The first part began by asking respondents how much AI collaboration they had at work. At present, there are relatively few authoritative scales on the application of AI, and those that exist are mainly scales at the macro-organization level. For example, patent citations data have enabled researchers to explore another channel of measuring the trends in technological progress (Kwon et al., 2017). Several studies measured the level of enterprise AI by the number of AI-related patent applications. The purpose of this study was to explore the strength of individual work AI collaboration from a micro perspective, rather than the degree of AI technology in the entire organization. Therefore, the study measured the AI collaboration variables through whether employees need to use AI at work and how often and how long they use it (Bhattacherjee and Sanford, 2006).
The second part of the survey was mainly about the measurement of job demand and job control in work. At present, the measurement of job demand and job control is mainly focused on Job Content Questionnaire (JCQ). Many studies have suggested that the measurement results of this scale are less affected by factors such as country, position, and so on, so it is relatively universal and can be used for job perception in multiple different countries and regions (Karasek, 1979; Niedhammer, 2002). This study mainly used the translated scale with good reliability and validity verified to measure job demand and job control variables. On this basis, appropriate deletion of irrelevant items was performed to fit the study. Richter's 7-level scale was used for all the scale items in this study, with 1–7 representing the respondents' perceived degree of satisfaction, from “very dissatisfied" to “very satisfied." The scale will not be described again here.
The third part of the questionnaire was mainly used to measure employees' sense self-efficacy. In the existing research on the field of professional work, the self-efficacy scale has been fully developed. This study was based on the professional self-efficacy scale developed by Betz and Hackett (2011). On this basis, we made appropriate deletions and modifications to make the scale more universal and keep 10 items.
The fourth part of the questionnaire addressed the Learning Goal Orientation (LGO) of the employees measured in the work. Learning goal orientation is mainly used to measure the extent to which employees think learning at work is important. At present, there is no relatively complete scale specially used to measure employees' learning orientation. Therefore, this study modified and changed the employee goal-oriented scale proposed by Yperen and Janssen (2002), extracted the part about learning goal-oriented, and finally formed a learning goal-oriented scale with six items.
The fifth part of the questionnaire measured employee learning behavior (LB) in organization. We summarized the self-learning ability and behavior of employees in organizations with complex, changing environments by looking at nine aspects: discovery behavior, invention behavior, selection behavior, implementation behavior, promotion behavior, reflection behavior, knowledge acquisition behavior, knowledge output behavior, and knowledge base building behavior. One or two questions addressed each issue, and 15 questions finally formed the measurement scale of employees' learning behavior in this study.
Finally, in the last section of the questionnaire, we designed items to capture respondents' personal demographic information, including age, gender, industry, department, position, and education level.
4.2 Sample and data collection
This research was divided into pre-investigation and formal investigation.
The pre-survey mainly aims to collect data on a small scale and test the reliability and validity of the questions. Unreasonable items are deleted and changed in advance to make the data collected in the large-scale survey more accurate. A total of 65 questionnaires were collected in the pre-survey stage. The results showed that JC2 and JC5 factor load coefficients did not meet the requirements. Thus, these two items were removed for following analyses.
Then, we conducted a large-scale questionnaire survey among employees of different genders, ages, and positions. To ensure the relevant of AI context and to reduce the sampling bias, the survey was only conducted in high-tech company including Baidu, Jingdong and Alibaba, in which AI collaboration is common. In addition to 65 small sample questionnaires, a total of 550 questionnaires were obtained, and 455 valid questionnaires were obtained, with an effective recovery rate of 82.83% after the ones with too-short response times or obvious invalidity were removed. The information distribution of samples is described in Table 1.
In terms of gender, there were more women, who accounted for 56.48% of respondents. In terms of age distribution, the sample was mainly distributed between those between 31 and 35 years old and those under 25 years old. Nearly three-quarters of the respondents were under 35 years old, which reflected that the sample selected in this study had a medium age structure and was in line with the age distribution of the labor force.
As far as the job characteristics of the sample, the respondents mainly worked in the internet/technology sector (26.59% of the total number), followed by finance (14.51%) and public utilities (11.65%). Most of the respondents were employed in R&D and sales departments in their companies, and most of them were employees of enterprises. A small number of them were senior managers.
In terms of education level, most of the subjects surveyed had a bachelor's degree or higher. These respondents accounted for about 65% of the total.
Comprehensive analysis showed that the distribution of the survey objects was relatively balanced. This can reduce the error caused by sample imbalance to a certain extent and basically meet the requirements of the survey, which can be further analyzed.
5. Results
Thinking that our analysis is concerned with testing a theoretical framework from a prediction perspective, and the structural model is more complex, including mediating variables and moderating variables, we used partial least squares (PLS) predict algorithm to test our hypotheses. PLS-SEM can handle these situations well and can work well with small or large sample sizes (Hair et al., 2019). The method has been developed by Shmueli et al. (2016). It uses training and holdout samples to generate and evaluate predictions from PLS path model estimations. The (fit) quality criteria are the most relevant output values for assessing both phases of the PLS-SEM estimation. Specifically, the criteria and generally suggested interpretation are summarized in Table 2.
5.1 Reliability and validity analysis
First, we analyzed the outer loadings of questionnaire questions, and the test results were as follows (Table 3):
We then conducted reliability and validity analyses on the collected questionnaire samples and conducted reliability and validity analysis on five variables, namely job demands, job control, self-efficacy, learning goal orientation and learning behavior.
It can be seen from Table 4 that the Cronbach α coefficients of the five variables all met the standard requirement of 0.7. This means that the variables showed good consistency internally and thus passed the reliability test. As shown in Table 5, the square roots of AVE of the five variables were 0.758, 0.776, 0.781, 0.881, and 0.806, respectively. These values were higher than the correlation coefficient between any two variables, which shows that the five variables had good discriminant validity.
5.2 Hypothesis testing
5.2.1 Test of the main model
Following the standard procedure of testing a structural model, we obtained the results. The results of R2 showed that the model accounted for 20.0%, 13.6 and 17.9% of the variances in continuance usage and repaired trust, respectively, indicating an acceptable level of explanatory power. The results of path coefficients showed that the five hypothesized paths were significant. We use the standardized residual mean root (SRMR) criterion to assess the model's goodness of fit. In this case, the SRMR was 0.061, less than 0.08. The results indicated an acceptable level of explanatory power and a satisfactory model fit.
As shown in Figure 3, we found job demand (β = 0.151, p < 0.001) and job control (β = 0.327, p < 0.001) were positively correlated with self-efficacy, and the positive correlation was significant. Thus, H1 and H2 were supported. There was a significant positive correlation between self-efficacy and learning goal orientation (β = 0.372, p < 0.001), which meant H3 was supported. Self-efficacy (β = 0.311, p < 0.001) and learning goal orientation (β = 0.201, p < 0.001) were significant positively correlated with employees' learning behavior. Therefore, H4 and H5 were supported.
5.2.2 Test of the mediating role of self-efficacy and learning goal orientation
We used the plug-in of SPSS Process Model4 to verify the mediating role of learning goal orientation between self-efficacy and employee learning behavior. It also verified the mediating effect of self-efficacy between job demand-control and learning goal orientation. The structure of the SPSS Process Model4 is shown in Figure 4.
The results are shown in Tables 6 and 7. It can be considered that the mediation model was established and partially mediated, and that H6 and H10 was supported.
5.2.3 Test of the moderating effect of AI collaboration
We added a moderator variable to the structural model of PLS to verify the moderating effects of AI collaboration. In PLS path modeling, we chose the two-stage method to construct the interaction term. Standardized data were then used for the calculation of the product terms of the interaction effect. The results of path coefficients show that two of the three moderating effects are significant.
As shown in Figure 5, we found collaboration with AI can positively moderate the positive impact of job demand on self-efficacy (β = 0.132, p < 0.01), while it cannot moderate the positive effect of job control (β = 0.026, p > 0.05). That is, H7 was supported while H8 was not supported. Collaboration with AI also had a significant positive moderating effect on the relationship between self-efficacy and learning behavior (β = 0.108, p < 0.05); therefore, H9 was supported.
In addition to the content mentioned earlier, we also explored other moderating effects of artificial intelligence (AI) collaboration. We used the SPSS Process Model 58 plugin to verify whether AI collaboration has a moderating mediating effect on learning goal orientation and self-efficacy. The research results show that AI collaboration did not exhibit any moderating mediating effects on learning goal orientation and self-efficacy.
6. Discussion
6.1 Summary of findings
So far, H1-H10 have been verified in this paper, and the results are shown in Table 8.
Based on the JDC model, this study has discussed the mechanism of employee learning behavior in the context of the application of AI at work, aiming to explain and predict employee learning behavior.
First, we have verified the positive effects of job demand and job control on employee self-efficacy. Next, through the model of self-efficacy – learning goal orientation – learning behavior, we have further proved that self-efficacy will have a positive effect on the mediating variable (employees' learning goal orientation), and the two will work together to have a positive effect on employees' learning behavior. Finally, the mediating role of self-efficacy between JDC and learning goal orientation was investigated.
This study also found that compared with the low level of AI collaboration, the positive impact of job demand on self-efficacy will be strengthened under a high level of AI collaboration, while the positive impact of job control will be only slightly strengthened, which proves the effect of artificial intelligence. In the digital age, the anxiety that the work brought by AI may be removed is not obvious. In more cases, AI can improve the efficiency and convenience of employees, reduce the triviality of employees' work, and reduce the pressure brought by work requirements.
At the same time, we found that the level of AI collaboration positively moderates the positive correlation between self-efficacy and employee learning behavior on the whole. In other words, the application of AI does contribute to the learning behavior of employees.
Another point worth noting is that, although no hypothesis was made, the beginning of the study explored the impact of demographic information on employees' self-efficacy and learning behavior through variance analysis. The interesting results show that the education level of employees has a significant positive correlation with self-efficacy and learning goal orientation, and the higher the education level, the stronger the self-efficacy of employees. However, they also show the learning behavior of employees will be influenced by more factors, and the positive impact of higher education will not run throughout the whole life of an individual.
6.2 Theoretical contribution
This study makes the following theoretical contributions.
This paper expands the traditional JDC model by introducing self-efficacy as the outcome of JDC model. We verify the positive and direct effects of job demand and job control on employee self-efficacy, which makes the theory richer. Previous studies mainly focused on the extension of the model, such as adding job tense factor (Bradley, 2007).
This paper also expands the theory of employee learning by revealing the formulation process of employee learning behavior. Based on the 3P model, we explore the positive relationship between self-efficacy and employee learning behavior, and explores the mediating role of learning goal orientation. It proves that self-efficacy can positively influence employees' learning behavior through the mediating variable of learning goal orientation, which provides the basis for the study of employees' learning behavior in organizations. Previous studies usually explore the antecedents of employee learning behavior from the perspective of students' self-efficacy and teachers' support (Ginns et al., 2014), while the formulation process of employee learning behavior is ignored.
This paper also enriches the JDC model and the theory of employee learning by introducing a context factor named AI collaboration which plays moderating roles in our model. Previous studies mainly examine employee learning behavior in traditional organization context, while the AI application creates an emerging context. Complex processes can be simplified in collaboration with AI collaboration, which change employees' workflows and learning environments. We explore the job demand – control – self-efficacy model and the role of self-efficacy and learning behavior model, finding that although the level of AI collaboration does not affect the relationship between job control and self-efficacy, it positively moderates the relationship between job demand and self-efficacy. Finally, we find that with a high-level AI collaboration, the positive influence of self-efficacy on learning behavior is enhanced.
6.3 Practical implications
Organizations need to cultivate excellent learning atmosphere at any time, encourage employees to accumulate knowledge and learn, and cultivate creative thinking to adapt to the organizational transformation in the era of artificial intelligence. Based on our findings, we put forward some guidelines for organizations that want to enhance their employees' learning behavior in the context of AI collaboration.
Organizations should specify reasonable job demand and job control intensity, and design reasonable job demand and job control. They should enhance employees' perceptions of both job demand and job control, since both have a positive effect on employee self-efficacy, and they should match. For example, organizations can try to create more relaxed environments and give employees more freedom to make decisions at work, which can enhance job control. At a high level of latitude regarding decisions, employees do not experience their jobs as stressful, despite those jobs being very psychologically demanding. Jobs with high demand and high control provide sufficient intrinsic motivation, and employees are open to accepting new challenges.
In order to develop, companies will have to deal with a situation of gradually increasing job demand. In the era of big data, organizations should introduce relevant artificial intelligence technology to improve the self-efficacy of employees, thus enabling them to learn better. We should not only manage the results of employee learning, but also strengthen the formation process of employee learning, in which self-efficacy and goal orientation play an important role.
Organizations need to manage AI collaboration in workplace. Learning should be emphasized in the organization, and employees should be wary of slack or non-learning behavior. In the context of AI Collaboration, AI can accomplish a lot of work, which may cause employees to give up learning. This study warns against this phenomenon. In addition, due to the positive moderating effect of AI collaboration on employee learning behavior, organizations should consider introducing AI technology into their work. Organizations can carry out AI technology teaching and training, so that employees can really apply AI technology to high demand-control jobs to achieve better collaborative effect.
Organizations can provide appropriate work resources and support, including necessary equipment, technology, and human resources to assist employees in completing their job tasks. Additionally, providing psychological health support and work/life balance measures is important. Encouraging employees to set clear learning goals and providing relevant training and development opportunities is also essential. By working with employees, their career development goals can be identified, and personalized training plans can be developed to achieve these goals. Providing appropriate feedback and rewards enhances employee motivation and goal orientation. This can include recognizing and praising employees' achievements at work and providing meaningful feedback and suggestions to help them improve. Utilizing artificial intelligence technology to provide personalized learning experience and support can improve employees' learning effectiveness by analyzing their learning styles and interests in order to offer suitable training content and reference materials. Organizations can develop appropriate career development plans and promotion paths for employees to understand their career prospects within the organization and provide related support and training opportunities to help them achieve their career goals. Cultivating a good learning culture and knowledge-sharing atmosphere is important. Organizations can encourage employees to share their experiences and knowledge, and provide suitable platforms and opportunities to promote exchange and learning among employees. Effective feedback mechanisms can be established to collect employees' feedback and suggestions on training and development opportunities and make timely improvements and optimization. In summary, to support employee learning and development, organizations need to take a series of measures including providing resources and support, encouraging goal setting, providing feedback and rewards, utilizing artificial intelligence technology, developing career development plans, cultivating a learning culture and knowledge-sharing atmosphere, and establishing effective feedback mechanisms. These measures will help enhance employees' learning motivation and self-efficacy, improve their learning abilities and promote their career development level.
6.4 Limitations and future research
The following limitation merits consideration and offers promising opportunities for future studies.
This study does not explore the impact of AI acceptance on employee behavior. Based on technology acceptance model (Davis and Davis, 1989), which describes employees' perception of emerging technologies, those they have the capability to use and perceive as easy to use will further affect their use tendencies. Self-efficacy has a certain effect on staff. The lack of a measurement of levels of self-efficacy means the experiment may not be very good individual test of artificial intelligence in the work of the utility.
AI has different impacts on people in different industries and departments, and with different skill requirements and positions, which may result in unemployment of low-skilled workers and salary increase of high-skilled workers. Therefore, it can be speculated that AI application has different effects on the behavior of employees with different skill levels. However, this study only distinguishes between job demands, not skill levels. Whether and how the degree of application of AI and the acceptance degree of new technology will affect the learning behavior of employees, as well as whether it will affect occupations with high and low skill requirements at different levels and in different directions, will be subjects to be studied in the future. The field of AI is still an emerging topic, and there are still more principles behind human–machine collaboration waiting to be explored and discovered.
7. Conclusion
Based on JDC model, this paper has discussed the mechanism of employee learning behavior under the background of the application of artificial intelligence technology in work, aiming to explain and predict employee learning behavior. In today's era of artificial intelligence, employees are more likely to produce positive self-efficacy in the context of high job demand-control matching, which has a positive correlation effect on employees' learning tendency, thus producing positive and serious learning behavior.
Starting from a new perspective, this paper has combined individual learning behavior with the JDC model, expanded the JDC model according to the background of the times, and provided a new direction for the following innovative research and development. At the same time, some suggestions have been put forward for enterprises, suggesting that enterprise leaders should pay attention to the sense of work control of employees while improving the job demand, effectively improve the level of self-efficacy and learning goal orientation of employees through technical cooperation and other means, and manage AI collaboration in work.
Figures
Figure 1
JDC working pressure model
Figure 2
Research model
Figure 3
Test results of main model
Figure 4
SPSS process Model4
Figure 5
Results of moderating effects
Sample description statistics
| Frequency | Percentage | ||
|---|---|---|---|
| Gender | Male | 198 | 43.52 |
| Female | 257 | 56.48 | |
| Age | Below 25 | 115 | 25.27 |
| 26–30 | 87 | 19.12 | |
| 31–35 | 131 | 28.79 | |
| 36–40 | 99 | 21.76 | |
| 41–45 | 13 | 2.86 | |
| 46–50 | 7 | 1.54 | |
| More than 50 | 3 | 0.66 | |
| Industry | Industrial | 39 | 8.57 |
| Financial | 66 | 14.51 | |
| Technology | 121 | 26.59 | |
| Consumer services | 46 | 10.11 | |
| Health care | 43 | 9.45 | |
| Telecommunications service | 35 | 7.69 | |
| Public utilities | 53 | 11.65 | |
| Others | 52 | 11.43 | |
| Department | Research and development | 99 | 21.76 |
| Purchasing | 38 | 8.35 | |
| Production | 44 | 9.67 | |
| Sales | 51 | 11.21 | |
| Personnel | 56 | 12.31 | |
| Finance | 46 | 10.11 | |
| Integrated management | 55 | 12.09 | |
| Logistics | 19 | 4.18 | |
| Others | 47 | 10.33 | |
| Position | Senior manager | 13 | 2.86 |
| Middle management | 63 | 13.85 | |
| Junior managers | 93 | 20.44 | |
| Enterprise staff | 286 | 62.86 | |
| Education | Below high school | 45 | 9.89 |
| Junior college | 70 | 15.38 | |
| Undergraduate | 267 | 58.68 | |
| Master | 68 | 14.95 |
Source(s): Author's work
Criteria for evaluating the quality of the measurement model
| Quality criteria | Description |
|---|---|
| R2 | Coefficient of determination. Indicates the percentage of variance of endogenous variables explained jointly by exogenous variables. Values between 0 and 1 can be taken, where values close to 0 indicate poor fit and values close to 1 indicate good fit |
| f2 | Measures effect size and the strength of the relationship between the variables on a numeric scale related to total, explained, and error variances
|
| Construct Reliability and Validity | Cronbach's alpha is a measure of internal consistency or reliability of a construct's measure—that is, how closely related the set of items comprising the construct are as a group. General guidelines on Cronbach's alpha for construct reliability and validity are
|
| Like Cronbach's Alpha, Composite reliability (CR) is a reliability indicator, but Cronbach's Alpha assumes factor loadings to be the same for all items, whereas CR takes into consideration the varying factor loadings of the items. Acceptable values of CR are generally considered 0.7 and above | |
| Average of variance extracted (AVE) is an indicator of convergent validity that measures the amount of variance that is captured by a construct in relation to the amount of variance due to measurement error. Generally, an AVE of at least 0.5 or higher is demanded; otherwise, the variance of error is more than the variance explained, which is considered unacceptable | |
| Discriminant validity determines whether the constructs in the model are highly correlated among themselves or not. It is generally suggested that the square root of AVE should be higher than the correlation of the construct with others | |
| Model fit | The standardized root mean square residual is an indicator of the average of the normalized residual between the observed covariance matrix and the assumed covariance matrix. Although recommendations vary from source to source, a good adjustment is generally considered to be less than 0.10 or 0.08 |
| Outer loadings | For reflection models, these are key indicators that show the path of latent variables to observed variables. Thus, they show how much each observable variable or item contributes in absolute terms to the definition of a constructed variable or potential variable. Normally the expected load value is greater than 0.6 |
| Outer weights | They are typical indicators for model formation because they illustrate the trajectory from observed variables to potential variables. These represent the relative contribution of indicators to the definition of their corresponding variables. It's also expected to be greater than 0.7 |
Source(s): Author's work
Test results of outer loadings
| JC | JD | LB | LGO | SE | |
|---|---|---|---|---|---|
| JC1 | 0.729 | ||||
| JC2 (dropped) | **0.147 | ||||
| JC3 | 0.782 | ||||
| JC4 | 0.79 | ||||
| JC5 (dropped) | **−0.235 | ||||
| JC6 | 0.728 | ||||
| JC7 | 0.716 | ||||
| JC8 | 0.772 | ||||
| JD1 | 0.827 | ||||
| JD2 | 0.848 | ||||
| JD3 | 0.645 | ||||
| JD4 | 0.776 | ||||
| JD5 | 0.768 | ||||
| LB1 | 0.765 | ||||
| LB10 | 0.809 | ||||
| LB11 | 0.815 | ||||
| LB12 | 0.84 | ||||
| LB13 | 0.808 | ||||
| LB14 | 0.744 | ||||
| LB15 | 0.76 | ||||
| LB2 | 0.746 | ||||
| LB3 | 0.769 | ||||
| LB4 | 0.705 | ||||
| LB5 | 0.776 | ||||
| LB6 | 0.806 | ||||
| LB7 | 0.811 | ||||
| LB8 | 0.783 | ||||
| LB9 | 0.764 | ||||
| LGO1 | 0.882 | ||||
| LGO2 | 0.89 | ||||
| LGO3 | 0.88 | ||||
| LGO4 | 0.891 | ||||
| LGO5 | 0.868 | ||||
| LGO6 | 0.874 | ||||
| SE1 | 0.841 | ||||
| SE10 | 0.787 | ||||
| SE2 | 0.821 | ||||
| SE3 | 0.843 | ||||
| SE4 | 0.812 | ||||
| SE5 | 0.793 | ||||
| SE6 | 0.821 | ||||
| SE7 | 0.779 | ||||
| SE8 | 0.751 | ||||
| SE9 | 0.806 |
Source(s): Author's work
Reliability and validity
| Cronbach's alpha | rho_A | Composite reliability | AVE | |
|---|---|---|---|---|
| JC | 0.852 | 0.856 | 0.89 | 0.575 |
| JD | 0.846 | 0.882 | 0.883 | 0.602 |
| LB | 0.954 | 0.961 | 0.959 | 0.61 |
| LGO | 0.942 | 0.944 | 0.954 | 0.776 |
| SE | 0.94 | 0.943 | 0.949 | 0.65 |
Source(s): Author's work
Pearson correlation and AVE square root value
| JC | JD | LB | LGO | SE | |
|---|---|---|---|---|---|
| JC | 0.758 | ||||
| JD | 0.287 | 0.776 | |||
| LB | 0.284 | 0.202 | 0.781 | ||
| LGO | 0.27 | 0.2 | 0.131 | 0.881 | |
| SE | 0.371 | 0.245 | 0.389 | 0.273 | 0.806 |
Source(s): Author's work
Mediating effect of learning goal orientation
| Object | A (SE=>LGO) | B (LGO=>LB) | a*b (Mediation) | a*b (95% BootCI) | c’ (Direct) | Result |
|---|---|---|---|---|---|---|
| SE=>LGO=>LB | 0.3974 | 0.1811 | 0.072 | 0.0288–0.1241 | 0.2962 | partial mediation |
Source(s): Author's work
Mediating effect of self-efficacy
| Object | A (JD/JC=>SE) | B (SE=>LGO) | a*b (Mediation) | a*b (95% BootCI) | c’ (Direct) | Result |
|---|---|---|---|---|---|---|
| JD=>SE=>LGO | 0.2073 | 0.3766 | 0.0781 | 0.0315–0.1434 | 0.0986 | partial mediation |
| JC=>SE=>LGO | 0.4534 | 0.3361 | 0.1524 | 0.0846–0.2433 | 0.1967 | partial mediation |
Source(s): Author's work
Hypothesis test results
| Hypothesis | ||
|---|---|---|
| H1 | The level of job demand will positively affect the self-efficacy of employees | Supported |
| H2 | The level of job control will positively affect the self-efficacy of employees | Supported |
| H3 | Employees' self-efficacy will positively affect their learning goal orientation | Supported |
| H4 | Employees' self-efficacy will positively affect their learning behavior | Supported |
| H5 | Employees' learning goal orientation will positively affect their learning behavior | Supported |
| H6 | Employees' learning goal orientation plays an intermediary role in the relationship between their self-efficacy and learning behavior | Supported |
| H7 | Collaboration with AI positively moderates the positive impact of job demand employees' self-efficacy | Supported |
| H8 | Collaboration with AI positively moderates the positive impact of job control on employees' self-efficacy | Not supported |
| H9 | Collaboration with AI positively moderates the positive impact of employees' self-efficacy on learning behavior | Supported |
| H10 | Employees' self-efficacy plays an intermediary role in the relationship between their job demand-control and learning goal orientation | Supported |
Source(s): Author's work
Appendix Measurement Items
Coordination between worker and AI (Bhattacherjee and Sanford, 2006)
AI1. Whether you need to use some kind of AI technology in your work.
AI2. The number of times a week that AI technology is used for collaborative work.
AI3. Time spent using AI technology for collaborative work.
Job demand (Karasek, 1979; Niedhammer, 2002)
JD1. Always need to get things done quickly.
JD2. Work hard to get the job done.
JD3. There is not enough time to finish the work.
JD4. Too many tasks at work.
JD5. Subject to conflicting demands from all sides.
Job Control (Karasek, 1979; Niedhammer, 2002)
JC1. Being asked to learn new things.
JC2. Very repetitive work.
JC3. Work requires creativity.
JD4. The job is technically demanding.
JC5. I have very little say in how the work is carried out.
JC6. At work, I can do a lot of different things.
JC7. Opinions are often influential.
JC8. The opportunity to develop your special talents.
Self-efficacy (Betz and Hackett, 2011)
OSES1. Although I have encountered many difficulties in the process of work, I am confident that I can do it well.
OSES2. Always confident in the face of new tasks.
OSES3. Believe that I can do the job well as required.
OSES4. Believe that your ability to work is no worse than others.
OSES5. When faced with a difficult problem, I can find several solutions.
OSES6. I am satisfied and enthusiastic about my current job.
OSES7. Personal promotion can be achieved.
OSES8. Read work-related books regularly and keep learning.
OSES9. Make frequent new suggestions for responsible work.
OSES10. Always full of energy at work.
Learning Goal Orientation (Yperen and Janssen, 2002)
LGO1. My work is successful when I am able to acquire new knowledge or skills that were difficult to master in the past.
LGO2. My work is successful when I learn something that motivates me to keep going.
LGO3. My work is successful when I learn new knowledge or skills through hard work.
LGO4. My job is successful when I learn something interesting and new at work.
LGO5. My work is successful when I learn something that makes me want to practice more.
LGO6. My work is successful when I acquire new knowledge or skills.
Learning Behavior
LB1. Early and accurate detection of all kinds of new changes and trends related to my work.
LB2. Identify potential problems, challenges or dangers in my work early and accurately.
LB3. Come up with new responses to changes in your job.
LB4. Be willing to offer new ideas.
LB5. Willing to propose creative measures.
LB6. Be able to make correct comparison, trade-offs and choices when faced with multiple considerations or options at work.
LB7. Can efficiently compare, choose, and make choices when faced with multiple considerations or alternatives.
LB8. Turn my work ideas into concrete actions.
LB9. Turn my work ideas into reality.
LB10. Apply my good work experience in many aspects of my work and benefit from it.
LB11. Learn lessons from mistakes in your work so that similar mistakes in my work do not happen again.
Be good at drawing regularly from what happened at my previous job.
LB13. Be good at acquiring knowledge/experience on the job through various channels (e.g. books, publications, websites, etc.).
LB14. Record and accumulate my thoughts, knowledge and experience.
LB15. Keep my knowledge and experience in good order, easy to keep and use.
Source(s): Author's work.
References
Ahmad, A., Cooper, C., Khan, Z., Golgeci, I. and Ali, I. (2021), “Artificial intelligence and human workers interaction at team level: a conceptual assessment of the challenges and potential HRM strategies”, International Journal of Manpower, Vol. 43, pp. 75-88, doi: 10.1108/IJM-01-2021-0052.
Akram, M.W., Mahar, S. and Ullah, M. (2017), “The influence of organization based self-esteem, role breadth self-efficacy and voluntary learning behavior on career success through mediation of perceived employability: a case of Pakistani banking”, Global Regional Review, Vol. 2, pp. 253-272.
Alam, A. (2022), “Employing adaptive learning and intelligent tutoring robots for virtual classrooms and smart campuses: reforming education in the age of artificial intelligence”, Advanced Computing and Intelligent Technologies, Vol. 914, pp. 395-406, doi: 10.1007/978-981-19-2980-9_32.
Baard, P.P., Deci, E.L. and Ryan, R.M. (2014), “Intrinsic need satisfaction: a motivational basis of performance and well-being in two work settings”, Journal of Applied Social Psychology, Vol. 34 No. 10, pp. 2045-2068, doi: 10.1111/j.1559-1816.2004.tb02690.x.
Bancans, A., Poell, R. and Meyers, C. (2014), “To what extent is the effect of personal characteristics (Self-efficacy, learning goal orientation and extraversion) on job satisfaction mediated by learning activities?”, Tilburg University.
Bandura, A. (1977), Self-efficacy: The Exercise of Control. W H Freeman/Times Books/ Henry Holt.
Bandura, A. (1986), Social Foundations of Thought and Action: A Social Cognitive Theory, Prentice Hall, p. 617.
Bandura, A. (1994), “Self-efficacy: the exercise of control”, doi: 10.1002/9780470479216.corpsy0836.
Bandura, A. and Wood, R. (1989), “Effect of perceived controllability and performance standards on self-regulation of complex decision making”, Journal of Personality and Social Psychology, Vol. 56, pp. 805-814, doi: 10.1037/0022-3514.56.5.805.
Betz, N.E. and Hackett, G. (2011), “Applications of self-efficacy theory to understanding career choice behavior”, Journal of Social and Clinical Psychology, Vol. 4 No. 3, pp. 279-289.
Bhattacherjee, A. and Sanford, C. (2006), “Influence processes for information technology acceptance: an elaboration likelihood model”, MIS Quarterly, Vol. 30 No. 4, pp. 805-825, doi: 10.2307/25148755.
Biggs, J. (1993a), “What do inventories of students' learning processes really measure? A theoretical review and clarification”, British Journal of Educational Psychology, Vol. 63 No. 1, pp. 3-19, doi: 10.1111/j.2044- 8279.1993.tb01038.x.
Biggs, J.B. (1993b), “From theory to practice: a cognitive systems approach”, Higher Education Research and Development, Vol. 12 No. 1, pp. 73-85, doi: 10.1080/0729436930120107.
Biggs, J. (2001), “Enhancing learning: a matter of style or approach?”, in Perspectives on Thinking, Learning, and Cognitive Styles, Lawrence Erlbaum Associates Publishers, pp. 73-102.
Bradley, G. (2007), “Job tenure as a moderator of stressor–strain relations: a comparison of experienced and new-start teachers”, Work and Stress, Vol. 21 No. 1, pp. 48-64, doi: 10.1080/02678370701264685.
Cao, D., Meadows, M., Wong, D. and Xia, S. (2021a), “Understanding consumers' social media engagement behaviour: an examination of the moderation effect of social media context”, Journal of Business Research, Vol. 122, pp. 835-846, doi: 10.1016/j.jbusres.2020.06.025.
Cao, G., Duan, Y., Edwards, J.S. and Dwivedi, Y.K. (2021b), “Understanding managers' attitudes and behavioral intentions towards using artificial intelligence for organizational decision-making”, Technovation, Vol. 106, 102312, doi: 10.1016/j.technovation.2021.102312.
Chowdhury, S., Budhwar, P., Dey, P.K., Joel-Edgar, S. and Abadie, A. (2022), “AI-employee collaboration and business performance: integrating knowledge-based view, socio-technical systems and organisational socialisation framework”, Journal of Business Research, Vol. 144, pp. 31-49, doi: 10.1016/j.jbusres.2022.01.069.
Chughtai, A.A. and Buckley, F. (2010), “Assessing the effects of organizational identification on in‐role job performance and learning behaviour”, Personnel Review, Vol. 39 No. 2, pp. 242-258, doi: 10.1108/00483481011017444.
Davis, F. and Davis, F. (1989), “Perceived usefulness, perceived ease of use, and user acceptance of information technology”, MIS Quarterly, Vol. 13, 319, doi: 10.2307/249008.
Deci, E.L. and Ryan, R.M. (1985), “Conceptualizations of intrinsic motivation and self-determination”, in Intrinsic Motivation and Self-Determination in Human Behavior, Springer, pp. 11-40.
Diallo, M.F., Diop-Sall, F., Djelassi, S. and Godefroit-Winkel, D. (2018), “How shopping mall service quality affects customer loyalty across developing countries: the moderation of the cultural context”, Journal of International Marketing, Vol. 26 No. 4, pp. 69-84, doi: 10.1177/1069031X18807473.
Doef, M. V.d. and Maes, S. (1999), “The Job Demand-Control (-Support) Model and psychological well-being: a review of 20 years of empirical research”, Work and Stress, Vol. 13 No. 2, pp. 87-114, doi: 10.1080/026783799296084.
Dunkin, M. and Biddle, B. (1974), The Study of Teaching, Holt, Rinehart and Wisnton, New York.
Dweck, C.S. (1986), “Motivational processes affecting learning”, American Psychologist, Vol. 41 No. 10, pp. 1040-1048, doi: 10.1037/0003-066X.41.10.1040.
Dweck, C.S. and Reppucci, N.D. (1973), “Learned helplessness and reinforcement responsibility in children”, Journal of Personality and Social Psychology, Vol. 25 No. 1, pp. 109-116, doi: 10.1037/h0034248.
Edmondson, A. (1999), “Psychological safety and learning behavior in work teams”, Administrative Science Quarterly, Vol. 44 No. 2, pp. 350-383.
Edmondson, A., Dillon, J. and Roloff, K. (2007), “Three perspectives on team learning: outcome improvement, task mastery, and group process”, Academy of Management Annals, Vol. 1 No. 1, pp. 269-314, doi: 10.5465/078559811.
Fryer, L.K. and Ginns, P. (2018), “A reciprocal test of perceptions of teaching quality and approaches to learning: a longitudinal examination of teaching-learning connections”, Educational Psychology, Vol. 38 No. 8, pp. 1032-1049, doi: 10.1080/01443410.2017.1403568.
Ginns, P., Martin, A.J. and Papworth, B. (2014), “Student Learning Theory goes (back) to (high) school”, Instructional Science, Vol. 42 No. 4, pp. 485-504, doi: 10.1007/s11251-013-9291-4.
Graham, J.D. and Bray, S.R. (2015), “Self-control strength depletion reduces self-efficacy and impairs exercise performance”, Journal of Sport and Exercise Psychology, Vol. 37 No. 5, pp. 477-488, doi: 10.1123/jsep.2015-0064.
Gurkok, H. and Nijholt, A. (2012), “Brain-computer interfaces for multimodal interaction: a survey and principles”, International Journal of Human-Computer Interaction, Vol. 28 No. 5, pp. 292-307, doi: 10.1080/10447318.2011.582022.
Häusser, J.A., Mojzisch, A., Niesel, M. and Schulz-Hardt, S. (2010), “Ten years on: a review of recent research on the job demand-control (-support) model and psychological well-being”, Work and Stress, Vol. 24 No. 1, pp. 1-35, doi: 10.1080/02678371003683747.
Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. (2019), “When to use and how to report the results of PLS-SEM”, European Business Review, Vol. 31 No. 1, pp. 2-24, doi: 10.1108/EBR-11-2018-0203.
Hirst, G., Knippenberg, D.V. and Zhou, J. (2009), “A cross-level perspective on employee creativity: goal orientation, team learning behavior, and individual creativity”, Academy of Management Journal, Vol. 52 No. 2, pp. 280-293.
Jarrahi, M.H. (2018), “Artificial intelligence and the future of work: human-AI symbiosis in organizational decision making”, Business Horizons, Vol. 61 No. 4, pp. 577-586, doi: 10.1016/j.bushor.2018.03.007.
Johnson, J.V. and Hall, E.M. (1988), “Job strain, work place social support, and cardiovascular disease: a cross-sectional study of a random sample of the Swedish working population”, American Journal of Public Health, Vol. 78 No. 10, pp. 1336-1342, doi: 10.2105/ajph.78.10.1336.
Karasek, R.A. (1979), “Job demands, job decision latitude, and mental strain: implications for job redesign”, Administrative Science Quarterly, Vol. 24 No. 2, pp. 285-308, doi: 10.2307/2392498.
Kevin, M. and Ana, F.I. (2019), “Case study - customer relation management, smart information systems and ethics”, The ORBIT Journal, Vol. 2 No. 2, pp. 1-24, doi: 10.29297/orbit.v2i2.114.
Kim, S. and Lee, J. (2015), “Employee self-concepts, voluntary learning behavior, and perceived employability”, Journal of Managerial Psychology, Vol. 30 No. 3, pp. 264-279.
Koo, B., Curtis, C. and Ryan, B. (2021), “Examining the impact of artificial intelligence on hotel employees through job insecurity perspectives”, International Journal of Hospitality Management, Vol. 95, 102763, doi: 10.1016/j.ijhm.2020.102763.
Kuleto, V., Ilic, M., Dumangiu, M., Rankovic, M., Martins, O.M.D., Paun, D. and Mihoreanu, L. (2021), “Exploring opportunities and challenges of artificial intelligence and machine learning in higher education institutions”, Sustainability, Vol. 13 No. 18, 10424, doi: 10.3390/su131810424.
Kwon, S., Lee, J. and Lee, S. (2017), “International trends in technological progress: evidence from patent citations, 1980-2011”, The Economic Journal, Vol. 127 No. 605, pp. F50-F70, doi: 10.1111/ecoj.12314.
Mikalef, P. and Gupta, M. (2021), “Artificial intelligence capability: conceptualization, measurement calibration, and empirical study on its impact on organizational creativity and firm performance”, Information and Management, Vol. 58 No. 3, 103434, doi: 10.1016/j.im.2021.103434.
Mirbabaie, M., Brunker, F., Mollmann, N.R.J. and Stieglitz, S. (2022), “The rise of artificial intelligence - understanding the AI identity threat at the workplace”, Electronic Markets, Vol. 32 No. 1, pp. 73-99, doi: 10.1007/s12525-021-00496-x.
Nam, T. (2019), “Technology usage, expected job sustainability, and perceived job insecurity”, Technological Forecasting and Social Change, Vol. 138, pp. 155-165, doi: 10.1016/j.techfore.2018.08.017.
Newman, A., Tse, H.H.M., Schwarz, G. and Nielsen, I. (2018), “The effects of employees' creative self-efficacy on innovative behavior: the role of entrepreneurial leadership”, Journal of Business Research, Vol. 89, pp. 1-9, doi: 10.1016/j.jbusres.2018.04.001.
Niedhammer, I. (2002), “Psychometric properties of the French version of the Karasek Job Content Questionnaire: a study of the scales of decision latitude, psychological demands, social support, and physical demands in the GAZEL cohort”, International Archives of Occupational and Environmental Health, Vol. 75 No. 3, pp. 129-144, doi: 10.1007/s004200100270.
Paschen, J., Wilson, M. and Ferreira, J.J. (2020), “Collaborative intelligence: how human and artificial intelligence create value along the B2B sales funnel”, Business Horizons, Vol. 63 No. 3, pp. 403-414, doi: 10.1016/j.bushor.2020.01.003.
Peng, Y.W. and Mao, C. (2015), “The impact of person-job fit on job satisfaction: the mediator role of self efficacy”, Social Indicators Research, Vol. 121 No. 3, pp. 805-813, doi: 10.1007/s11205-014-0659-x.
Perry-Smith, J.E. (2006), “Social yet creative: the role of social relationships in facilitating individual creativity”, Academy of Management Journal, Vol. 49 No. 1, pp. 85-101, doi: 10.5465/amj.2006.20785503.
Salanova, M., Peiró, J.M. and Schaufeli, W.B. (2002), “Self-efficacy specificity and burnout among information technology workers: an extension of the job demand-control model”, European Journal of Work and Organizational Psychology, Vol. 11 No. 1, pp. 1-25, doi: 10.1080/13594320143000735.
Schilling, J. and Kluge, A. (2009), “Barriers to organizational learning: an integration of theory and research”, International Journal of Management Reviews, Vol. 11 No. 3, pp. 337-360, doi: 10.1111/j.1468-2370.2008.00242.x.
Shmueli, G., Ray, S., Estrada, J. and Chatla, S.B. (2016), “The elephant in the room: predictive performance of PLS models”, Journal of Business Research, Vol. 69 No. 10, pp. 4552-4564.
Shrestha, Y.R., Ben-Menahem, S.M. and Krogh, G.V. (2019), “Organizational decision-making structures in the age of artificial intelligence”, California Management Review, Vol. 61 No. 4, pp. 66-83.
Sima, V., Gheorghe, I.G., Subic, J. and Nancu, D. (2020), “Influences of the industry 4.0 revolution on the human capital development and consumer behavior: a systematic review”, Sustainability, Vol. 12 No. 10, p. 4035, doi: 10.3390/su12104035.
Sowa, K., Przegalinska, A. and Ciechanowski, L. (2021), “Cobots in knowledge work: human – AI collaboration in managerial professions”, Journal of Business Research, Vol. 125, pp. 135-142, doi: 10.1016/j.jbusres.2020.11.038.
van der Maas, H.L.J., Snoek, L. and Stevenson, C.E. (2021), “How much intelligence is there in artificial intelligence? A 2020 update”, Intelligence, Vol. 87, 101548, doi: 10.1016/j.intell.2021.101548.
Vincent, V.U. (2021), “Integrating intuition and artificial intelligence in organizational decision-making”, Business Horizons, Vol. 64 No. 4, pp. 425-438, doi: 10.1016/j.bushor.2021.02.008.
Yoon, J. and Kayes, D.C. (2016), “Employees' self-efficacy and perception of individual learning in teams: the cross-level moderating role of team-learning behavior”, Journal of Organizational Behavior, Vol. 37 No. 7, pp. 1044-1060, doi: 10.1002/job.2092.
Yperen, N.W.V. and Janssen, O. (2002), “Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands”, Academy of Management Journal, Vol. 45 No. 6, pp. 1161-1171, doi: 10.5465/3069431.
Acknowledgements
Funding: This work was supported jointly by the grant of National Natural Science Foundation of China (72231004 and 72271182), the grant of National Social Science Foundation of China (18ZDA109), and the funding from Shandong Industrial Internet Innovation and Entrepreneurship Community.
Corresponding author
About the authors
Aihui Chen is an associate professor of College of Management and Economics, Tianjin University in China. His research focuses on AI, AIOT, and AI-powered organizations. His research has been published in Journal of Management Information Systems, Journal of Information Technology, Information & Management, Journal of Business Research, Journal of Global Information Management, International Journal of Information Management, and others.
Tuo Yang is an undergraduate student of College of Management and Economics, Tianjin University in China. She majors in Information Management and Information Systems. Her research focuses on artificial intelligence.
Jinfeng Ma is a graduate student of College of Management and Economics, Tianjin University in China. She majors in Management Science and Engineering. Her research focuses on digital-intelligent oppression.
Yaobin Lu is a professor of Management Science and Information Management at School of Management, Huazhong University of Science and Technology. He is the Director of Research Center of Information Management. He is a PI for seven national scientific grants. He was a visiting professor at MISRC, University of Minnesota, USA. He is an associate editor of Information & Management. He published six books in Information Management, and more than 60 papers in peer-reviewed journals like MIS Quarterly, Journal of Management Information Systems, Journal of Information Technology, Information Systems Journal, Decision Support Systems, International Journal of Information Management, Information & Management, Journal of Business Research, and International Journal of Human-Computer Interaction. From 2014 to 2022, he is listed as Chinese Most Cited Researchers by Elsevier. His papers have academic influence and high citation according to the statistics of Elsevier.