The next speakers in this session at the 5th Data Donation Symposium at the Weizenbaum-Institut in Berlin are Alexandra Wölfle and Michaela Forrai, whose interest is in linking panel data with chatbot history data donations. Their focus is on emerging adults, aged 18-29 – these are now also a key demographic for chatbot use, with some two thirds likely to have used a chatbot, for a range of purposes including informational and instrumental, social support, personal identity, and entertainment and escapism use.
Such chatbot use may foster wellbeing especially in professional contexts; but it may also have detrimental effects especially for particularly heavy chatbot users. This can be understood through self-determination theory, which suggests that wellbeing is connected to a person’s sense of autonomy, competence, and relatedness. These needs may also motivate individuals to adapt their behaviour, while media use may satisfy these needs.
The need for autonomy may motivate chatbot use, for instance; meanwhile, behavioural control may affect changes in chatbot use, especially also for everyday advice. Chatbot use may also foster personal reflection processes, and help in the identification of personal strengths and weaknesses; it may support autonomy satisfaction, and encourage users to try new activities.
Further, intrinsic motivation to acquire knowledge may motivate use; this is true both across practical everyday uses as well as broader knowledge contexts. This may also link to the use of chatbots for personal feedback, examples, and exercises, which build confidence and enable users to achieve academic goals; overall, this increases competence satisfaction.
Meanwhile, users reporting loneliness use chatbots as a source of socialisation and companionship, as chatbots are perceived as more accepting and less judgmental; this leads to a greater sense of social connectedness and reduced loneliness, while however possibly also reducing real-world socialisation.
The project seeks to test this through a two-wave survey of German-speaking emerging adults, which also connects data donations of the five most recent chatbot conversations. As data downloads from chatbots take some time, some further workarounds for these delays had to be built into the process here; there are also additional challenges around the download package format and potentially problematic information included in such donations. How to reflect the results of the analysis back to the donors also remains an open question.
So far, the preliminary analysis has not yet found any clear motivational patterns that explain chatbot use, other than from an autonomy need to informational uses of chatbots; this may be related to the as yet small sample of users, though, or the selection of conversations users have chosen to share.
