Talking to a Chatbot: AI Recommendations and Belief Updating

Job Market Paper

Abstract: This study investigates whether AI-powered chatbots are more persuasive than static sources of information. In a lab experiment, participants reported their beliefs about the correct answer to a series of logical reasoning questions both before and after receiving recommendations generated by the same AI model. In one treatment, participants received real-time recommendations from a chatbot, while in the other, they received static, pre-generated recommendations. The informational content of the recommendations, including both the recommended answers and the reasoning behind them, was distributionally matched across treatments. Participants who interacted with the chatbot were slightly more likely to update their beliefs following the AI recommendations, with a 31.46 percentage point change compared to 26.31 in the static treatment (p = 0.084). Further analyses reveal that the chatbot was especially persuasive among participants whose initial beliefs clashed with the recommendations.

Monitoring as a Service

Abstract: This study explores whether a monitoring service can help mitigate procrastination and improve commitment. In a real-effort experiment, participants needed to complete 80 tasks across three sessions over three weeks. In the first week, after making a non-binding plan for how many tasks to complete during the final two weeks, they were randomly assigned to either a Control or a Monitored group. In the Monitored group, participants were connected with a Monitor via WhatsApp, who would observe their progress and send out reminders if they delayed their work. Monitoring improved adherence to the initial plan by 12.75 percentage points (p < 0.05) and completing all the required tasks by 19.47 percentage points (p < 0.001), mainly by reducing dropout rate. The findings suggest that combining external oversight with performance-related reminders can be an effective method to combat procrastination.

Emotional Discrimination

(with Sigrid Suetens and Boris van Leeuwen)

Abstract: While ethnic discrimination is still a major concern, not much is known about the role of emotions on discrimination itself. We conducted an online experiment where native Dutch participants, playing the responder role, received a low offer from a proposer with either a native Dutch name or an ethnic minority name, and had to decide whether to accept or reject this low offer. Their emotions could be triggered in some treatments where they actually experienced the low offer before making a choice. We found that the responders did not discriminate nor react emotionally, as we did not find significant treatment effects on the rejection rate to the low offer.