@thesis{Ambacher2024, author = {Ambacher, Julian Emanuel}, title = {Designing a user-friendly and optimized version of a user interface for a large language model (LLM)}, publisher = {Technische Hochschule Ingolstadt}, address = {Ingolstadt}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:573-51556}, pages = {IX, 134}, school = {Technische Hochschule Ingolstadt}, year = {2024}, abstract = {This thesis focuses on the creation of a user-friendly and efficient user interface (UI) for a large language model (LLM). As LLMs become more widely used in different fields, it is important to have a UI that meets user needs when interacting with them [1], [2], [3]. The main goal of this study is to find key design principles, design guidelines, and methods that ensure that the UI is easy to use, efficient, and satisfying. To achieve this, the thesis analyzes the interaction principles according to ISO 9241-110 on the existing LLM tool Chat GPT and focuses also on other different UI designs [4]. For this purpose, a prototype is created based on the analysis. This thesis executes a usability test to evaluate the usability of the initial prototype. It focuses on factors including flexibility, accuracy of system answers, ease of navigation, and feature efficacy. The feedback from this test is used to enhance the design. The second iteration of the prototype is evaluated using several measures, including the System Usability Scale (SUS) and the User Experience Questionnaire (UEQ). In addition, feedback was gathered from four experts in the field of UX and AI. The results showed improvements in ease of use, task efficiency, and overall user satisfaction. The results highlight the importance of clear guidance, privacy, and system responsiveness. In summary, the thesis provides a set of design guidelines and recommendations for creating a user-friendly user interface for LLMs. These include the aspects of user-centered design, contextual help features, and optimized workflows. In addition, it provides guidance for the future development of AI tools that prioritize both functionality and a positive user experience.}, language = {en} }