College Students’ Trust in Artificial Intelligence for Advice on Big Life Decisions

Authors: Joelle Sacks, William Hayes, Elaine Yu, Trevor Rosenthal

SUNY Campus: Binghamton University

Presentation Type: Poster

Location: Old Union Hall

Presentation #: 25

Timeslot: Session A 9:00-10:00 AM

Abstract: The accessibility of artificial intelligence (AI) technologies such as ChatGPT has opened the door to numerous possibilities for its use, including giving advice. AI chatbots may be consulted for advice on difficult life decisions, but to what extent would people trust AI for various kinds of advice? In this exploratory study, 182 undergraduates completed a survey gauging trust in AI chatbots, message boards (e.g., Reddit), and other people for advice on 46 life decisions from seven major categories. Participants were asked whether or not they had faced the decisions and follow-up questions about what type of advice they considered and from which sources. Participants were then asked to rate their level of trust in the 3 different sources of advice on a 100-point sliding scale. We found that participants generally trusted other people the most, online message boards the second most, and AI chatbots the least. For decisions involving family or relationships, participants were less likely to trust AI chatbots and online message boards when compared to other types of decisions. The results of this study can inform future research on trust in artificial intelligence and its potential role in decision-making. Future studies can be done to examine individual differences to determine if any variables are strong predictors of trust in AI.