Conrad Liszka, Cesar Sandoval, Dominique Davis, Haley Narita, Jayden Ramirez de Arellano, & Dr. Weldon Zane Smith
The field of generative deep-learning models, colloquially referred to as “artificial intelligence” or “AI”, has rapidly grown in recent years, both in terms of available services and societal impact (Li, 2023). However, comparatively little psychology-based research has been performed in relation to generative AI, especially regarding the most current services available. The purpose of the current study is to expand the available research on how a person’s psychology may interact with their perception of AI, specifically version 3.5 of ChatGPT. Previous research suggests that AI involvement affects people’s opinions across various contexts; listeners seem to like pieces of music less when they believe those pieces are composed by an AI (Shank et al., 2023). Additionally, the knowledge that an AI will be involved in interview evaluations for certain jobs appears to negatively affect motivation to apply for or pursue them (Mirowska, 2020). Thus, the current research seeks to further investigate the relationship between perceived AI involvement and opinions regarding them and their usage. To achieve this, a deception was performed in which certain participants were told that the survey they were currently taking had either been completely written by AI or constructed with the assistance of AI. Edited screenshots of responses from ChatGPT were included as evidence to aid in the deception. A control group composed of participants who do not experience any deception was also included. Afterward, opinions on AI were collected from all groups using an adapted version of the General Attitudes Towards Robots Scale with 20 items (Koverola et al., 2022). A one-way between groups ANOVA was used to analyze the responses of 1982 participants. There was no significant difference in overall opinions regarding AI across levels of perceived AI involvement, which may suggest an emerging acceptance or apathy toward broader AI usage. These findings can be useful for individuals and organizations that aim to use artificial intelligence, especially in cases necessitating disclosure of its assistance. Practical implications of the current research extend across creative, social, and industrial/organizational contexts, which will be further explored in the presentation.