Sense of Autonomy Enhances Consumer Decision Making with AI

Retailers increasingly use AI-based recommendation systems to help consumers make purchase decisions, but could these shopping gurus be too good at telling us what to do?

A new study by Guiyang Xiong, professor of marketing, and three colleagues, forthcoming in MIS Quarterly, suggests that the key to making these AI systems more effective lies in restoring a sense of autonomy—the feeling that users are in control of the process. Across two field experiments and several lab studies, the researchers found that consumers who feel that AI is assisting rather than determining choices for them are more likely to make purchases and less likely to return them later.

“AI-based systems are really good at offering optimal choices, but these are based on the average consumer, not the unique preferences and priorities of the individual user,” says Xiong, who co-authored the paper with Jove Jinghui Hou, assistant professor of information systems at the University of Houston’s Bauer College of Business; Shuai Yang, professor at Donghua University’s Glorious Sun School of Business and Management; and Paul Pavlou, dean of the Patti and Allan Herbert Business School and Leonard M. Miller University Chair Professor at the University of Miami. “Paradoxically, when AI takes over their decision making, people may lose some sense of autonomy and end up making less satisfying decisions.”

The study tested this idea in a fashion retail store using an AI-powered apparel shopping assistant, a floor mirror with a 55-inch touchscreen. The mirror scans customers from head to toe and recommends items based on factors like body shape, skin tone, size and seasonal trends. When store messaging emphasized that the AI was simply offering help—not taking over—shoppers bought more and felt more satisfied with their decisions. Their sense of autonomy was also greater when they used the recommendation system on their smartphones rather than on store-owned devices.

Controlled lab experiments supported the field findings, shedding light on the underlying psychological mechanisms and ruling out alternative explanations.

The results have practical implications for any retailer looking to adopt AI. The researchers caution against the current practice of portraying AI as a professional advisor or all-knowing expert. Instead, they recommend putting consumers back in the driver’s seat, with better outcomes for both businesses and buyers.

Xiong, G. (2025), Enhancing Consumer Decision-Making with AI: The Role of the Sense of Autonomy (with Hou, J.J., Yang, S. and Pavlou, P.), MIS Quarterly.