Abstract
The rapid advancement of Artificial Intelligence (AI), particularly Large Language Models (LLMs), has transformed how users interact with technology by enabling high-levellanguage interpretation and generation. However, the text-based interaction design commonly employed poses challenges for novice users and those with physical constraints,
necessitating proficiency in prompt engineering to achieve satisfactory results for complex queries. This thesis addresses the usability and accessibility issues inherent in AIintegrated applications by exploring alternative input modalities to enhance the user
experience. Through the development and evaluation of a recipe generator application,
this research investigates how diverse user interface elements such as sliders, buttons,
and tags can improve the usability and efficiency of interactions. A user study involving
participants with varying levels of expertise assesses the impact of these modalities on
user engagement, satisfaction, and performance. The findings highlight the potential of
tailored User Interface (UI) elements to lower the entry barrier for LLM usage, offering
insights into future AI application designs that prioritize inclusivity and ease of use.
| Date of Award | 2024 |
|---|---|
| Original language | English (American) |
| Supervisor | Kathrin Probst (Supervisor) |