청소년의 또래집단 규범이 비윤리적 행동 중단의도, 인공지능 챗봇에 대한 친밀감과 상호작용 만족도에 미치는 영향 : 메시지 프레임의 조절효과

Effects of Adolescents’ Peer Group Norms on Intention to Stop Unethical Behaviors, Closeness to and Satisfaction with AI Chatbots : The Moderating Role of Message Framing

Given that the interactions between humans and artificial intelligence (AI), particularly the use of AI chatbots by adolescents, are increasing, it is necessary to discuss related ethical issues. For instance, adolescents are highly likely to be exposed to socially undesirable language use online, and AI chatbots can easily learn users’ language use. Thus, adolescents’ use of AI chatbots may lead to ethically problematic issues. However, empirical studies that have investigated ethical issues in the context of adolescents and AI chatbot interactions are limited. The present study examined how the interactions between adolescents and AI chatbots were conducted when they were in unethical situations, and how the persuasive effects of AI chatbots worked in such situations. Specifically, the present study examined how adolescents" peer group norms and message framing influenced their intention to stop unethical behaviors, their sense of closeness to AI chatbots, and their satisfaction with AI chatbots. A 3 (peer group norms toward an unethical behavior: agree vs. disagree vs. no specific opinion) x 2 (message framing: positive vs. negative) factorial design was utilized in two separate online experiments with South Korean adolescents aged 13-18. The scenarios presented to the participants involved unethical situations such as online course auto-enrollment or unauthorized absenteeism, thereby simulating real-life experiences that adolescents may frequently encounter. Results showed that closeness to an AI chatbot was highest when the AI chatbot had an opposite opinion to peer group norms. However, intention to stop unethical behaviors and satisfaction with the AI chatbot based on peer group norms did not present significant differences. In terms of message framing, negative framing in the AI chatbot’s communication led to higher levels of both chatbot closeness and interaction satisfaction with the AI chatbot. In addition, the interaction effect between peer group norms toward the unethical behaviors and AI chatbot message framing was not significant. The importance of the present study lies in its empirical examination of how AI chatbots can positively affect adolescents’ moral reasoning and attitude development from an ethical standpoint. By demonstrating the nuanced ways in which AI chatbots can shape the ethical landscapes of young minds, the study contributes to understanding of adolescents-AI chatbot interaction. It also suggests the potential of AI chatbots in ethical education, providing vital clues on how AI chatbots can effectively enhance ethical values and attitudes among adolescents. The present study also offers practical implications in the area of educational AI chatbot, offering possible ways to employ AI chatbots for promoting ethical education.

본 연구는 청소년의 또래집단 규범과 인공지능 챗봇 메시지 프레임이 비윤리적 행동 중단의도, 인공지능 챗봇에 대한 친밀감, 챗봇과의 상호작용에 대한 만족도에 미치는 영향을 분석함으로써 비윤리적 상황에서 인공지능 챗봇의 효과성을 탐구했다. 이를 위해 3(비윤리적 행동에 대한 또래집단 규범: 찬성 vs. 반대 vs. 특별한 의견 없음) x 2(메시지 프레임: 긍정 vs. 부정)의 개체 간 요인설계를 바탕으로 만 13-18세 청소년을 대상으로 두 가지 온라인 실험을 수행하였다. 각 실험의 참여자는 여섯 개의 실험조건 중 하나에 무작위로 배치되었다. 분석 결과, 비윤리적 행동에 대한 또래집단 규범이 반대일 경우, 챗봇에 대한 친밀감이 높게 나타났으나 비윤리적 행동 중단의도와 챗봇과의 상호작용에 대한 만족도는 유의미한 차이를 보이지 않았다. 메시지 프레임의 경우, 부정 프레임 메시지에서 챗봇에 대한 친밀감과 상호작용에 대한 만족도가 높았다. 한편, 비윤리적 행동에 대한 또래집단 규범과 인공지능 챗봇 메시지 프레임의 상호작용 효과는 유의미하지 않은 것으로 나타났다. 본 연구는 윤리적 관점에서 청소년의 도덕적 의사결정과 태도 형성에 인공지능 챗봇이 긍정적인 영향을 미칠 수 있음을 실증적으로 살펴보았다는 점에서 의의가 있다. 또한, 윤리적 가치와 태도 교육에 있어 인공지능 챗봇의 활용 가능성을 시사한다.