The prompt-response paradox: A cybernetic perspective of generative artificial intelligence usage and corner-cutting behavior

The emergence and proliferation of generative artificial intelligence (GenAI) have transformed workplace interactions, reshaping many of them into dense “Prompt-response (P-R)” cycles characterized by a structural imbalance between the effort employees invest in crafting prompts and the quality of AI-generated feedback. While existing GenAI literature has largely focused on the scale, frequency, and operational patterns of its application, it has largely overlooked the inherent imbalances within these P-R cycles. Building on control theory, this study investigates how and when GenAI usage is related to employees’ corner-cutting behavior (CCB). Using a scenario-based experiment and a three-wave field survey, we demonstrate that: (i) performance pressure mediates the nonlinear relationship between GenAI usage and CCB, such that this nonlinear indirect effect shifts from positive to negative as GenAI usage intensifies; (ii) workplace status moderates the inverted U-shaped relationship between GenAI usage and performance pressure. Specifically, high status sharpens this curve, whereas low status flattens it, with a complete reversal to a U-shaped relationship occurring only under conditions of extremely low status. By conceptualizing GenAI usage through the dual lenses of expectation induction and cognitive friction operating within P-R cycles, our findings extend and refine traditional perspectives on human-AI interaction in the workplace.