A pilot study of human-AI conversational interaction and its impact on loneliness and wellbeing

With the growing accessibility of advanced artificial intelligence (AI) chatbots, there is a need to understand their impact on users' psychological wellbeing. This pilot study aimed to explore the subjective experiences of human-AI conversational interaction and its potential relationship with loneliness and life satisfaction.

A mixed-methods study was conducted among 19 psychology students who voluntarily chose one of two forms of support, i.e., interaction with a chatbot (N = 9 for the quantitative component; N = 11 for the qualitative component) or a conversation with a psychologist (N = 10 for the quantitative component; N = 9 for the qualitative component). The Satisfaction with Life Scale (SWLS) and the De Jong Gierveld Loneliness Scale were used. Qualitative data were collected through participant-maintained diaries and subjected to a thematic analysis.

The quantitative analysis revealed that the group that chose to interact with the chatbot was characterized by a higher sense of loneliness (M = 6.69) and lower life satisfaction (M = 20.67) compared to the group that chose contact with a psychologist (M = 3.66 and M = 24.20, respectively). The qualitative analysis revealed that the participants perceived the chatbot to be a safe and non-judgmental space for expressing emotions, while also recognizing its formulaic nature and limitations.

The findings of this pilot study suggest that interactions with AI may be particularly appealing to individuals experiencing greater emotional loneliness. The feasibility of the research procedure was also confirmed. Further research on larger and more diverse samples is necessary to verify these preliminary observations and to investigate the long-term effects of human-AI interaction.