This study investigates how an employee’s use of artificial intelligence (AI) impacts managers' subjective performance evaluations, particularly focusing on how managers' perceptions of AI influence their performance expectations and subsequent evaluations of their employees. Using an experiment manipulating AI reliability perceptions (high vs. low) and performance outcomes (success vs. failure), I predict and find that managers expect higher performance from employees using AI perceived as reliable. Further, consistent with my prediction, I find greater discrepancies between managers' expectations and evaluations when an employee uses unreliable AI and exhibits successful performance. Additionally, successful performance is more often attributed to the employee when AI is seen as unreliable, whereas no significant attributional differences were found for unsuccessful performance. These findings highlight the importance of managers' perceptions of AI in shaping managerial expectations and evaluations.