Shun Ye, George Mason University – AI Tools and Human Decision Making

AI may help us make decisions, but are they the right decisions?

Shun Ye, associate professor of information systems and operations management at George Mason University, examines the outcomes.

Shun Ye is an Associate professor of Information Systems and Operations Management at George Mason University’s Costello College of Business. With a Ph.D. in Management Information Systems from the University of Maryland, Dr. Ye’s research interests focus on digital platforms and open innovation. Published in top leading journals like MIS Quarterly and Information Systems Research, Dr. Ye engages in various professional organizations, such as the Association for Information Systems (AIS) and INFORMS. A multiple award winner for his teaching excellence, including the George Mason University Teaching Excellence Award, he teaches undergraduate, MBA, and MS Business Analytics programs.

AI Tools and Human Decision Making

AI-driven tools and features on online platforms aim to simplify and speed up consumer decision-making. However, my research shows that while this approach appears effective for locating information more quickly, it might unintentionally lead consumers to make poorer choices.

For our study, my co-authors and I used data from two widely used restaurant review platforms, Yelp and TripAdvisor, to assess the impact of Yelp’s AI-powered image categorization feature. This feature categorizes restaurant review images into five sections – food, drinks, menu, inside, and outside. The idea is to make it easier to find relevant restaurant images, so that users can choose restaurants more efficiently. For our experiment, we compared Yelp’s data, with TripAdvisor’s data, matching corresponding restaurants across 10 U.S. cities. These comparisons allowed us to track how Yelp’s AI feature changed actual consumer decisions, by analyzing ratings before and after the AI was added. We saw that while the AI photo categorizer appeared to cut down search time, it also led consumers to overlook relevant information that is often better captured in written reviews – like service quality. As a result, the caliber of decision-making declined.

We also found that Yelp’s AI feature caused what is called cognitive miser behavior, where consumers favor less mentally taxing information, such as AI images, over more complex or in-depth content – like written reviews. So, while review images can show off to consumers the tangible – cozy ambiance and mouthwatering food – this feature misses the mark on the intangible and led to an increase in consumer complaints related to service.

To make the most of these findings, platform managers and business owners who wish to implement AI tools on their platforms should consider complementing these features with existing tools, such as written reviews, in order to balance convenience and comprehensive decision-making.

Read More:

Email: sye2@gmu.edu
Phone: (703) 993-1844

Relevant Publications:

• Lianlian Jiang, Shun Ye, Liang Zhao, and Bin Gu. “ Do Reductions in Search Costs for Partial Information on Online Platforms Lead to Better Consumer Decisions? Evidence of Cognitive Miser Behavior from a Natural Experiment,” Information Systems Research, forthcoming
• Pallab Sanyal, and Shun Ye. “An Examination of the Dynamics of Crowdsourcing Contests: Role of Feedback Type,” Information Systems Research (35:1), 2024, pp. 394-413.
• Raveesh Mayya, Shun Ye, Siva Viswanathan, and Rajshree Agarwal. “Who Forgoes Screening in Online Markets and Why? Evidence from Airbnb,” MIS Quarterly (45:4), 2021, pp.1745-1776.