Emanuel Georg Schlosser
2019 unpublished
This research deals with the impact of the transparency factor on the evaluation of human-chatbot communication and the chatbot as a communication partner. In a qualitative research framework, a one-factorial Wizard-of-Oz experiment was set up (treatment: transparency factor), in which 12 subjects had to solve three tasks via Instant Messenger, either together with a Chatbot (Chatbot Anna) or a human (Anna). Including a service call, a small talk situation and the desert survival problem. The
more » ... rvey was then carried out by means of a questionnaire and qualitative interviews. The work provides a basic insight into the field of chatbot communication and shows that the transparency factor, ie the knowledge about the identity of the communication partner, has a decisive influence on the evaluation of the communication partner Chatbot, as well as the conversation itself , In the evaluation, two main emphases could be identified: human-to-human conversations are evaluated primarily on character and relationship level; while human-to-chatbot conversations are evaluated from an efficiency and understanding perspective (task & communication level). In addition, it could be shown that even different behavioral and evaluation patterns are evoked by the user alone or by chatbot. From the findings, it is finally possible to derive practice-relevant implications for the targeted use of chatbots.
doi:10.25365/thesis.59156 fatcat:hwr2su2gdvaw7mpibjox4xo6sm