Does human-AI collaboration promote or hinder employees’ safety performance? A job demands-resources perspective

Abstract

With the widespread implementation of artificial intelligence (AI), human-AI collaboration has become an important and influential employment model. However, there is no consensus in the literature on its effectiveness, and little is known about how it affects employees’ safety performance. Drawing on the job demands-resources (JD-R) theory, this study proposes a double-edged sword effect model of human-AI collaboration on employees’ safety performance. Using three-wave data from 286 employees, the findings show that human-AI collaboration fosters intrinsic motivation and work engagement, thereby enhancing safety performance (the motivation pathway), while also increasing workplace loneliness and job burnout, which inhibits safety performance (the strain pathway). Furthermore, employee resilience amplified the positive impact of human-AI collaboration on intrinsic motivation and mitigated its effects on workplace loneliness. These findings enrich our understanding of human-AI collaboration and offer practical insights for AI adoption and safety management in organizations.