关键词: AI驱动的工作非常规化, 基于AI工作重塑, 工作意义感, 工作特征模型

Abstract: As artificial intelligence (AI) increasingly serves as an automation tool that takes over routine work in organizations, employees are left with a greater proportion of non-routine tasks. This shift gives rise to AI-enabled job non-routinization (AIJNR). Prior research has mainly understood AIJNR as a change in job content, whereby employees move from routine tasks to non-routine tasks, and has examined its implications for creativity. However, AIJNR is not merely a process of making job content more non-routine. It may also restructure the value attributes and boundary structure of employees’ work, requiring employees to reconsider the meaning of their work. Drawing on the Job Characteristics Model (JCM), we propose that AIJNR influences work meaningfulness through two opposing pathways. On the one hand, when AI absorbs repetitive and rule-based tasks, employees may devote more attention to work that is consequential, judgment-intensive, and valuable for the organization, thereby enhancing perceived task significance and, in turn, work meaningfulness. On the other hand, AI may also fragment workflows and reduce employees’ opportunities to complete an identifiable whole piece of work, thereby undermining perceived task identity and, in turn, work meaningfulness. In addition, we examine AI-based crafting as a moderator that captures employees’ proactive adjustment to AI-enabled changes in work. We tested these ideas through three studies. Study 1 used archival collaboration data from the GitHub repository huggingface/transformers, including 286,773 event logs and 31,037 cases, to provide descriptive evidence of increasing job non-routinization after the diffusion of AI tools. User-level analyses based on 1,891 collaborators further indicated that AIJNR became more salient in the post-2022 period. Focusing on 764 collaborators in 2023–2024, we found that AIJNR was positively associated with indicators of task significance and negatively associated with indicators of task identity. Study 2 employed a video-based scenario experiment with 180 participants and showed that higher AIJNR increased task significance but decreased task identity, which in turn produced opposite indirect effects on work meaningfulness. Study 3, a three-wave field survey of 224 employees, replicated this dual-path pattern in a field setting. Moreover, AI-based crafting strengthened the positive relationship between AIJNR and task significance, whereas its moderating effect on the relationship between AIJNR and task identity was not significant. This research makes three theoretical contributions. First, it extends the job characteristics model by showing that AIJNR can serve as an important source of changes in core job characteristics. Second, it reveals the boundary role of AI-based crafting in shaping how AIJNR is translated into employees’ perceptions of job characteristics. Third, it incorporates technology-driven work structure change into research on the sources of work meaningfulness, offering a new perspective for understanding work design and employees’ meaning experience in the AI era.

Key words: AI-enabled job non-routinization, AI-based crafting, work meaningfulness, job characteristics model