Factors Influencing Individuals’ Intentions to Misuse LLM-Based Generative AI in Video Content Creation

Misusing generative AI in video content creation, such as creating deepfakes, leads to significant social losses and even crime. While individual users can become easily involved in generating social losses and committing crimes, scant research examines the factors influencing their intention to misuse generative AI in video content creation. To fill this gap, this study focuses on LLM-based GAI capable of manipulating linguistic information in videos, aiming to identify the key factors influencing individuals’ intentions to misuse generative AI, as well as their resistance to such technology. An online survey was conducted in June 2023, and data were collected from 400 respondents in Korea. The analyses identified associations of perceived misuse risk with misuse intention for AI voice and of favorability with resistance for AI human. Additionally, the mediation analyses identified significant indirect associations involving attitude and subjective norm in the relationships among risk perception, favorability, misuse intention, and resistance. These findings suggest that perceived misuse risk and attitudes toward disinformation creation are relevant correlates of misuse intention, rather than establishing that changing these perceptions would reduce misuse intention. Notably, the observed associations showed different descriptive patterns across media modalities. Furthermore, it highlights the relevance of perceived misuse risk when examining intentions to misuse LLM-based GAI.