How do Input modes affect politeness?

Read full text Summarise with VIRAT
Authors: Chanmi Lee, Felix Braun, Fabian Aleksic
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Large Language Models have undergone a huge boom over the past years, and are being included in more products every day. They are starting to supplant many traditional app-based approaches, and quickly becoming a major entry-point into people's computing devices. This results in a shift towards using natural language as an input method, rather than just using buttons and menus. Users have diverse ways of interacting with their devices, such as full-size hardware keyboards on notebooks, to touchscreen keyboards of varying sizes, all the way to speech input on smartwatches or via headphones. These different input modes already require some optimizations to enable applications that require input to work well, such as predefined answers for incoming messages on smartwatches. With the relevance of politeness already demonstrated, we want to build on this existing research and investigate the effect different input methods have on this. We believe that on top of an individual's baseline politeness (resulting both from character as well as mental model there's a bias introduced by the throughput of the input method used to communicate with the chatbot. Input methods such as smartwatch displays require considerably longer time to input the same number of words, making it more expensive to use pleasantries and other formalities not directly required by the query. Using smaller screens for typing also results in quicker fatigue and less comfort. We suspect that this results in a more direct, but also less polite interaction, compounded by (potential) frustration at the difficulty in getting the point across. A very common scenario for using smart assistants is the kitchen, where people want help with following recipes, and asking questions about what they need to do next. We're going to be simulating this scenario for our study.

Fields of Study & Subjects

Speech and dialogue systems AI in Service Interactions Language Discourse Communication Strategies

Publication details

Publication date
1 January 2026
Language
EN
Document type
Preprint