The Influence of Chatbot Design on User Reactions to Unsatisfactory Answers

Wednesday - April 10 2024, 22:41 UTC - 9 months ago

A study led by Professor Alfred Brendel from TU Dresden explored the influence of chatbot design on user reactions to unsatisfactory answers. The results showed that while human-like design may increase user satisfaction, it does not prevent aggression towards chatbots. Developers should carefully consider the positive and negative effects of human-like design elements.

Chatbots have become a popular tool for internet users seeking quick and efficient answers to their questions and requests. However, the reactions to chatbots have not always been positive, according to Professor Alfred Benedikt Brendel from TU Dresden. In a study led by Brendel and his international research team, it was hypothesized that the design of chatbots may have an influence on user responses to unsatisfactory answers. The results of the study yielded interesting findings and have important implications for the development and design of chatbots in the future.

Brendel explains that if a chatbot provides incorrect or confusing answers, it can lead to aggression from users towards the digital conversation partner. This aggression can range from verbal abuse to negative attitudes towards the provider or website itself. To test their hypothesis, the researchers conducted experiments with two different types of chatbots – one with human attributes and one without.

The human chatbot used in the study was equipped with a name, gender, and picture, and responded to questions in a friendly manner with appropriate emojis. On the other hand, the neutral chatbot had no such design elements. The results of the study showed that the human chatbot generally increased user satisfaction and reduced frustration. However, when faced with unsatisfactory answers, participants still exhibited aggressive behavior towards both the human and neutral chatbot. Overall, approximately 10% of users showed aggressive behavior towards the virtual assistants.

Interestingly, the intensity of aggressive behavior was reduced when interacting with the human chatbot compared to the neutral one. This suggests that human-like design elements may have a positive impact on user reactions, but should be carefully considered. According to Brendel, this has important implications for chatbot developers who should approach human-like design with caution and consider both positive and negative effects it may have on user behavior and attitudes.

In conclusion, the study provides valuable insight into the influence of chatbot design on user reactions to unsatisfactory answers. The use of chatbots is expected to continue to grow, making it increasingly important for developers to consider the implications of their design choices. Further research is needed to fully understand the impact of chatbot design on user behavior and attitudes.