University of Twente Student Theses
Quality of Interaction With Chatbots as a Function of Fairness, Appearance and Congruence With Participants' Declared Sex
Assen, L.J. (2024) Quality of Interaction With Chatbots as a Function of Fairness, Appearance and Congruence With Participants' Declared Sex.
|
PDF
2MB |
| Abstract: | The availability of knowledge through the information age is of great benefit to anyone. However, it also enables the spread of unfair information i.e., biased and incorrect information. It is known that unfair information may spread through artificial intelligence-driven chatbots (CA) due to the hallucination effect. However, no previous work has been done to understand its influence on the human-CA interaction. The present study focuses on this. Separately, the influence of the CA’s appearance and congruence were also investigated. This led to two problems that were explored; i) whether the CA’s appearance and congruence influence the flagging of information; and ii) whether the difference between the user’s expected and experienced quality of interaction differs across levels of (un)fairness and whether the CA’s appearance and congruence have an influence. A first study focussed on creating a set of clearly unfair stimuli. The second study focussed on the experimental assessment of the human-CA interactions. For this, a pre-post, between-subjects design was used. The results showed an increase in the quality of interaction in the completely fair condition and a slight decrease in the completely unfair condition, whilst no difference was found in the 50% unfair condition. The results also suggest that people expect CAs to be somewhat unfair. Lastly, no differences due to appearance or congruence were found. |
| Item Type: | Essay (Bachelor) |
| Faculty: | BMS: Behavioural, Management and Social Sciences |
| Subject: | 77 psychology |
| Programme: | Psychology BSc (56604) |
| Link to this item: | https://purl.utwente.nl/essays/100688 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page