Against the backdrop of artificial intelligence (AI) technology continuously integrating into organizational operations and job practices, traditional motivation theories, constrained by static research assumptions, struggle to reasonably explain the complex situation where employees experience both increased and decreased motivation. Rooted in motivational psychology and integrating Expectancy Theory with Self-Determination Theory, this paper constructs a dual-path motivation analysis framework. Focusing on two core characteristics—the capability paradox and relationship reconfiguration—the paper systematically explores the dynamic evolution patterns and intrinsic mechanisms of employee motivation in AI application scenarios. Empirical analysis based on multi-time-point tracking data reveals that AI integration into work settings primarily influences individual motivation systems through two paths: task reshaping and cognitive reshaping, resulting in differentiated impacts. High cognitive demand jobs commonly exhibit a capability paradox where “objective performance improves while subjective competence perception declines”, a contradiction significantly alleviated by employees’ AI literacy. Meanwhile, employees’ perceptions of work relationships are undergoing profound transformations, shifting from traditional interpersonal emotional support orientation to a rational orientation with clearly defined rights and responsibilities, with increased collaboration frequency failing to synchronously enhance relationship security. From a dynamic perspective, this study expands the applicability boundaries of classical motivation theories, providing robust theoretical support and practical references for organizations to optimize motivation management, refine incentive systems, and assist employees in better adapting to career transitions in AI-enabled environments.