Motivating employee performance improvement through negative feedback is a persistent challenge in performance management. As digital and intelligent technologies advance, some tech firms have begun leveraging cutting-edge tools to optimize performance management processes and enhance employee experience. Adopting a motives attribution perspective, this study compares the differentiated performance incentive mechanisms of artificial intelligence (AI) and human leaders when delivering negative feedback. Empirical findings show that, relative to human leaders, AI-provided negative feedback elicits stronger attributed performance-promotion motives and weaker attributed injury-initiation motives, which in turn lead to higher employee performance. Further grounding the inquiry in the indigenous Chinese context reveals the moderating role of leadership style: AI exhibits a greater advantage over authoritarian leadership in transmitting the indirect effects of negative feedback on performance through these two motives; however, compared with benevolent leadership, this differentiated indirect effect is attenuated. The study uncovers the performance-incentive effects of negative feedback in human-AI interaction, broadens the context, perspectives, and approaches for research on AI and negative feedback, and offers insights for the digital and intelligent transformation of performance management practices in Chinese enterprises.