Examining Factors Influencing Social Media User's Attitudes Toward AI-Generated Content

Abstract—The fast development of AI-generated content (AIGC) on social media platforms is challenging users to distinguish between human-made and AI-generated content. This study explores the factors influencing users' attitudes towards AIGC using the Partial Least Squares Structural Equation Modelling (PLS-SEM) approach. Based on the Technology Acceptance Model (TAM) and Cognitive Bias Theory, a conceptual model was developed with seven constructs in two paths: (1) Ethics of AI and Self-Awareness as predictors of AI Literacy, which influences Perceived Usefulness and Attitude; and (2) Perceived Risk influencing Trust, which then influences Attitude towards AIGC. Data were obtained via an online questionnaire distributed to social media users in Indonesia using purposeful sampling. From the 500 questionnaires distributed from April 2026 onwards, 450 responses were deemed valid. The tested model includes seven hypotheses with five latently exogenous variables and two endogenously latent variables. The results showed that self-awareness was the strongest predictor of AI literacy, while ethics of AI was not significant. Perceived Risk was a strong predictor of Trust, and Perceived Usefulness was the main determinant of Attitude, followed by Trust. AI literacy does not directly influence attitude, but rather through the mediation of perceived usefulness. These findings contribute to the literature on AI literacy and emphasise the importance of self-awareness and perceived risk in shaping users' responses to AIGC. 

Keywords—Al-generated content, Al literacy, perceived usefulness, trust, PLS-SEM, self- awareness