This thesis examines how different user interface (UI) designs influence user trust in generativeArtificial Intelligence (AI) tools. We conducted an experiment employing the Wizard of Ozmethod to test the trust level of three tools with different UI variations of ChatGPT. Ninevolunteer university students from diverse disciplines participated. We used questionnaires toassess participants' perception of trust after interacting with each tool, and after interacting withall tools. Results indicate that the trust levels among participants are influenced by generativeAI's UI design, particularly avatar design and text font. Most participants ranked ChatGPT as themost trustworthy compared to other tools despite sharing the same text source. The results alsohighlight the importance of conversational interfaces in building trust with generative AIsystems, with participants expressing a preference for interfaces that facilitate natural andengaging interactions. The study underscores the significant impact of UI on trust and aims toencourage more cautious trust in generative AIs.