More to Lose: The Adverse Effect of High Performance Ranking on Employees’ Preimplementation Attitudes Toward the Integration of Powerful AI Aids

Despite the growing availability of algorithm-augmented work, algorithm aversion is prevalent among employees, hindering successful implementations of powerful artificial intelligence (AI) aids. Applying a social comparison perspective, this article examines the adverse effect of employees’ high performance ranking on their preimplementation attitudes toward the integration of powerful AI aids within their area of advantage. Five studies, using a weight estimation simulation (Studies 1–3), recall of actual job tasks (Study 4), and a workplace scenario (Study 5), provided consistent causal evidence for this effect by manipulating performance ranking (performance advantage compared with peers versus no advantage). Studies 3–4 revealed that this effect was driven in part by employees’ perceived potential loss of standing compared with peers, a novel social-based mechanism complementing the extant explanation operating via one’s confidence in own (versus AI) ability. Stronger causal evidence for this mechanism was provided in Study 5 using a “moderation-of-process” design. It showed that the adverse effect of high performance ranking on preimplementation AI attitudes was reversed when bolstering the stability of future performance rankings (presumably counteracting one’s concern with potential loss of standing). Finally, pointing to the power of symbolic threats, this adverse effect was evident both in the absence of financial incentives for high performance (Study 1) and in various incentive-based settings (Studies 2–3). Implications for understanding and managing high performers’ aversion toward the integration of powerful algorithmic aids are discussed.