With an increase in interactions with non-human agents, understanding and forecasting the consequences of human interactions with them have become increasingly crucial. The present study focused on a subset of non-human agents known as virtual characters. To investigate whether people respond in the same way to a virtual character as they do to a real human, we replicated a classic test of human-human social interaction, specifically, the social facilitation and inhibition effects. The theory of social facilitation/inhibition states that in the presence of others, people perform simple or learned tasks better and complex or novel tasks worse. In the present study, participants were given different tasks (letter typing, Sudoku, and maze) of varying difficulty (easy and difficult) to complete either alone, in the presence of a real person, or in the presence of a virtual character. The participants' task performance was determined by the number of subjects completed within the specified time. The results of 60 participants demonstrated a significant improvement in the performance of easy tasks when a virtual character was present compared with performing those tasks alone, providing evidence for social facilitation effects. For the difficult tasks, task performance did not show decline in the presence of a virtual character. According to these findings, it is possible to avoid social inhibition in the presence of a virtual character. The present study will be a reference for scholars interested in further investigating how non-human agents impact social facilitation effects and implement virtual agents in educational settings.