◎ Editor|Digital Economy Pioneer
◎Source|Scientific and technological progress and countermeasures
◎Author|Zhang Heng, Gao Gaohua, Li Huiling
With the development of artificial intelligence technology, a large number of AI technologies have appeared in the workplace, bringing many challenges to traditional human resource management. On the one hand, companies want to introduce AI technology to replace human resources, but they are worried that some of AI's capabilities are not as good as humans. On the other hand, employees enjoy the convenience brought by AI technology, but also face the risk of being replaced by AI.
So does the application of AI technology have a gain effect or a loss effect on employees' innovation ability? This paper will focus on the impact of AI technology on employees' innovative behavior, and set up a "work requirements - resource model" to explore the moderating role of AI technology in the workplace to play a "double-edged sword" effect.
Theoretical hypothesis
The theoretical basis of this study is the work requirements-resource model (JD-R model). The model divides the work characteristics into two dimensions: work requirements and work resources. Job requirements include factors that consume physical and psychological costs, such as job insecurity, time pressure, etc.; Work resources include factors that promote growth and development, such as peer support, perception of work autonomy, etc. The JD-R model also emphasizes the "two-path" hypothesis that work has a gain path and a loss path, i.e., work resources positively affect individuals by increasing work input, while job requirements negatively affect individuals by increasing burnout.
The impact of workplace AI technology applications on employees' innovative behavior can be explained by two paths: job requirements and work resources. On the one hand, AI technology replaces some programmatic work, increasing employees' job insecurity and leading employees to reduce innovative behavior. On the other hand, AI technology reduces the workload of employees, increases the perception of work autonomy, and enables employees to have more time and energy to participate in innovative behaviors.
In addition, an individual's level of control over work affects employees' perception of job requirements and work resources. Among them, learning goal orientation means that individuals are willing to improve new skills and knowledge, and can effectively control the use process and results of AI with high work control. This research focuses on the boundary conditions of the "double-edged sword" effect of learning goal orientation in the application of AI technology on employees' innovative behavior.
Here's why: First, learning goal orientation determines how individuals process and view information in challenging AI situations. Individuals with high learning goal orientation have a higher degree of control over their work, and believe that the application of AI technology gives employees the opportunity to challenge themselves to improve themselves, and when individuals perceive that external motivation is consistent with self-preference and promote goal achievement, they will spontaneously transform external motivation into internal motivation, so as to regard the application of AI technology as an improving work resource rather than an obstructive work requirement;
Second, when conducting self-assessment, high-learning goal-oriented individuals pay special attention to whether the current work state has significantly grown and improved compared with the past. Therefore, in the application scenario of AI technology, when employees with high learning goals find that their skills are far from the real workplace, they will independently learn and make up for the lack of digital knowledge and skills, thereby increasing work resources and reducing work requirements.
1.2 AI Technology Application and Employee Innovation Behavior: The Loss Path Through Job Insecurity
According to the JD-R model, job insecurity is a typical job requirement, and drastic changes in the application of AI technology and technological advances are likely to trigger employee job insecurity. When employees perceive that their work safety feels threatened but cannot respond effectively, it will stimulate the individual's self-protection mechanism, which in turn will reduce the innovative behavior of employees.
Therefore, we put forward two hypotheses: the H1 hypothesis is that the application of AI technology is positively correlated with job insecurity, and the H2 hypothesis is that the application of AI technology negatively affects employees' innovative behavior through job insecurity.
1.3 AI Technology Application and Employee Innovation Behavior: Gain Path Perceived through Work Autonomy
Perception of work autonomy refers to the degree to which employees are able to exercise discretion when it comes to their work. This study predicts that the application of AI technology will stimulate employees' perception of work autonomy. On the one hand, AI technology is able to handle complex and high-cognitive work tasks, providing employees with timely and useful information and reducing their work burden. This helps provide idle resources and increase work flexibility, allowing employees to organize their own work processes and increase the perception of work autonomy. On the other hand, AI technology can continuously acquire and understand a large amount of data, and employees can freely arrange their working hours according to their needs, learn and apply new skills independently, optimize work procedures, and increase the perception of work autonomy. Based on these views, the H3 hypothesis is proposed: the application of AI technology is positively correlated with the perception of work autonomy.
According to the JD-R model, the perception of work autonomy as a work resource can stimulate the positive work state of the individual. With a sense of work autonomy, employees can learn and explore new knowledge and skills on their own, and invest more attention and resources to generate and implement new ideas, thereby increasing employees' innovative behavior. In addition, in AI application scenarios, employees mainly undertake creative, social and interpersonal related work. Employees with a high sense of work autonomy can make full use of idle resources brought by AI, actively communicate with colleagues and leaders, and try to use new methods and technologies to improve work processes and promote innovative behavior. Therefore, the H4 hypothesis is proposed: AI technology applications positively influence employees' innovative behavior through work autonomy perception. These findings have important implications for companies to understand how to stimulate employees' autonomous perception and promote innovative behavior through the application of AI technology.
1.4 The moderating role of learning goal orientation
An individual's level of control over work affects his or her perception of work resources and job requirements. Learning goal orientation refers to the attitude and tendency of individuals to learn new knowledge and master new skills to improve their own abilities, and can control their work and skill application at a high level. This study argues that learning goal orientation can mitigate the negative impact of AI technology application on job insecurity. First of all, individuals with high learning goal orientation pay attention to learning and progress, are willing to face challenging tasks, and strive to learn skills and professional knowledge in AI scenarios, so as to obtain the satisfaction of job competence and reduce job insecurity. Second, individuals with high learning goals-oriented have an advantage in competing with AI, are able to solve problems in the field of work, alleviate the worry of being replaced by AI, and reduce job insecurity. Conversely, individuals with low learning goal orientation do not focus on learning and developing new knowledge and skills, they believe that completing basic work tasks and avoiding risks are the most important, and the learning and challenges brought about by the application of AI technology do not match their needs, causing fear of technology replacing jobs and increasing job insecurity. Based on these views, the H5 hypothesis is proposed: learning goal-oriented negative moderation of the relationship between AI technology application and job insecurity.
In addition, combined with the previous research hypotheses H1 and H2, the mediating H6 hypothesis with moderation is further proposed: learning goal orientation plays a moderating role in the indirect effect of AI technology application in reducing employees' innovative behavior through job insecurity.
This study also suggests that learning goal orientation enhances the positive impact of AI technology application on the perception of work autonomy. Individuals with high learning goal orientation have the willingness to learn independently, can actively protect and use the idle resources brought by AI in dealing with daily work, independently arrange work plans and times, and actively learn and apply new knowledge and technologies, so as to perceive more work autonomy. On the contrary, individuals with low learning goal orientation do not pay enough attention to the improvement of self-ability and the acquisition of work resources, and will negatively view the work challenge of AI technology application and reduce the perception of work autonomy. Based on this, the H7 hypothesis is proposed: the relationship between learning goal-oriented positive regulation AI technology application and work autonomy perception.
In addition, combined with the previous research hypotheses H3 and H4, the mediating hypothesis H8 with moderation is further proposed: learning goal orientation plays a moderating role in the indirect effect of AI technology application to increase employees' innovative behavior through work autonomy perception.
Research methods
Study 1 recruited 300 full-time working subjects to participate in the experiment through the Credamo Seeing Platform, and after removing the unqualified samples, 278 valid samples were obtained, and 43.17% of the valid samples were male; The age concentration is under 30 years old, accounting for 64.75%.
Study 1 adopts 2 (AI technology application: high vs. low) (learning goal orientation: high vs. low) two-factor group design. In this study, participants were randomly assigned to four experimental scenarios. After the experiment began, the participants first filled in demographic variables, and in order to distinguish AI technology in the workplace from traditional technology (computers, Internet use, word processing or spreadsheet software, etc.), we first showed them the definition of AI technology applications. Immediately afterwards, the participants were invited to substitute their own role settings in the material as much as possible (assuming that they were Zhang San) and read a situational material. Participants then complete questionnaires based on the material, including manipulation tests, job insecurity, perception of work autonomy, and innovative behavior.
Manipulation of AI technology applications: adapted from the definition of AI technology applications and manipulation materials of predecessors (control group in parentheses): "Your ABC company has introduced a large number of (did not introduce) AI technology and equipment, and all aspects of work, such as reasoning, decision-making and problem solving, are mainly completed by these smart devices autonomously (mainly by you), and you often use (usually do not use) AI technology and equipment when implementing some work tasks. And spend (and don't spend) a lot of time working with AI technology."
Learning goal-oriented manipulation: Based on experimental materials developed by predecessors (control group in parentheses): "After graduating from college, you were hired by ABC. In the first six months, the company department has multiple projects working on at the same time, you actively choose a difficult and challenging (low) project, insist on learning new knowledge and acquiring new skills every day to improve your abilities, and you attach great importance (and you do not value) to try and explore, often (rarely) use most of your personal time to learn and create, eager (not eager) to learn from tasks and improve professional skills by studying hard".
Study results
Study 1: AI technology application was taken as the independent variable, learning goal orientation was taken as the moderating variable, gender and age were used as covariates, and job insecurity and work autonomy perception were taken as dependent variables. Analysis of variance shows that the main effect of AI technology application on job insecurity is significant. The H1 hypothesis holds.
At the same time, the main effect of AI technology application on work autonomy perception was significant, indicating that compared with the low AI technology application group, the high AI technology application group experienced a stronger perception of work autonomy, and the H3 hypothesis was true.
The ANOVA results show that the interaction between AI technology application and learning goal orientation on job insecurity is significant. As shown in Figure 2, the more AI technology is applied under low learning goal orientation, the more job insecurity increases, and the increase exceeds that of high learning goal oriented group, H5 hypothesis is true.
As shown in Figure 3, the more AI technology is applied under high learning goal orientation, the more work autonomy perception increases, and the increase significantly exceeds that of the low learning goal orientation group, and the H7 hypothesis is true.
Drawing on the Bootstrap test, the juxtaposition mediating roles of job insecurity and work autonomy perception in AI=technology application and employees' innovative behavior are examined, and the results are shown in Table 1. The results show that the assumptions H2 and H4 are true.
In this paper, 2000 analyses of the mediated mediating effect were extracted from the Bootstrap test, and the results are shown in Table 2, H6 and hypothetical H8 hold.
The above results provide preliminary evidence for the "double-edged sword" effect of AI technology application on employees' innovative behavior, and examine the key moderating mechanism of learning goal orientation. In order to ensure the validity of the research results, we also designed a questionnaire survey to re-test the overall model, which complemented the scenario experiment and jointly improved the reliability of the research results.
Study 2: Full-model questionnaire survey
Employees from four AI-applied companies in Beijing and Shenzhen were selected as survey subjects and paper questionnaires were distributed, covering four industries: healthcare, finance, intelligent manufacturing, and transportation. 418 valid questionnaires were obtained.
The results shown in Table 5 show that the application of AI technology is positively correlated with job insecurity and work autonomy perception, work insecurity is negatively correlated with innovative behavior, and work autonomy perception is positively correlated with innovative behavior.
Conclusion and discussion
Based on the JD-R model, this paper deeply discusses the "double-edged sword" effect of AI technology application in the workplace on employees' innovative behavior, including the loss path caused by job insecurity and the gain path formed by work autonomy perception, as well as the boundary conditions affecting the two effects. Through the two research designs of scenario experiment and questionnaire survey, it is concluded that the application of AI technology can negatively affect employees' innovative behavior by increasing job insecurity, and can also positively affect employees' innovative behavior by enhancing the perception of work autonomy. In addition, the enhancement of learning goal orientation will weaken the loss path of AI technology application and strengthen the gain path of AI technology application. This study not only broadens the research scope of AI technology application in organizational behavior, but also provides theoretical guidance and practical enlightenment for organizations to promote employees' innovative behavior in AI scenarios.
Specific advice to managers is as follows: guide employees correctly and encourage a rational view of AI technology as an opportunity rather than a threat. Help employees realize that AI can reduce work burden, increase work flexibility, and stimulate work autonomy and innovative behavior. Give employees emotional care and psychological safety, pay attention to employees' psychological fluctuations and emotional feelings, formulate employee help plans, and reduce the threat of AI to employees' work status and psychological safety. Advocate individual learning goal orientation to adapt to work changes in AI scenarios, enhance learning ability, cope with opportunities and challenges brought by AI, and stimulate work enthusiasm and innovation motivation. The learning goal orientation is used as the recruitment and selection criteria, and the employees with strong learning goal orientation are identified through personality characteristic tests. Conduct regular digital training, build a learning organization, guide employees to establish learning goal orientation, improve professional knowledge and digital literacy.
Due to space limitations, this article has been partially abridged. Original source: [1] Zhang Heng, Gao Gao Hua, Li Huiling. Gain or Loss: The "Double-edged Sword" Effect of Artificial Intelligence Technology Application on Employees' Innovative Behavior[J/OL].Scientific and technological progress and countermeasures:1-11.)
About us
"Digital Economy Pioneer" is a digital economy research and exchange platform under Chengdu Digital Lianhe Service Technology Co., Ltd. Focusing on digital application fields such as digital industry, digital infrastructure, digital governance, and digital ecology, reveal and record the bits and contexts of the development of the digital economy.
Digital Co-production Service is a digital economy industry think tank and industrial big data service provider, with full-process big data governance-analysis-decision-making support service capabilities, providing industrial economic development solutions and comprehensive services based on big data for governments at all levels and industrial operation institutions.