Revealing the dark and bright sides of ChatGPT: an exploratory study on user perceptions

• Main Authors: NAH, Fiona Fui-hoon, CAI, Jingyuan, ZHENG, Ruilin, CHEN, Langtao
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2023
• Subjects: ChatGPT; Human-conversational agent interaction; User media post analysis; Categorical Data Analysis; Graphics and Human Computer Interfaces
• Online Access: https://ink.library.smu.edu.sg/sis_research/9915
• Institution: Singapore Management University

A state-of-the-art AI conversational agent, ChatGPT, has attracted overwhelming attention and interest. ChatGPT draws on existing knowledge repositories and demonstrates excellent comprehension and content-generation capabilities. To better understand the potential and limitations of human–conversational agent interaction, we conducted an exploratory qualitative content analysis of user posts available on the social media platform, Reddit. Manifestations of the dark side of ChatGPT include algorithm and product limitations, the possibility of singularity, risks of job displacement, threats to information privacy and security, access and content restrictions, violation of content policy, addictive tendencies, and inappropriate use for cheating and plagiarism. The bright side of ChatGPT includes assistance with tasks, exploratory use, and value creation. This research highlights the importance of understanding user perceptions of AI conversational agents that can be used to guide the design, implementation, and regulation of generative AI products.