Lamita Farah

Shifting Personas explores how AI’s human-like persona and embodiment shape user engagement, behaviour, and perception in everyday interactions, such as those with virtual assistants, chatbots, and AI companions. At the centre of this project is an adaptive AI chatbot that tailors its responses based on the user’s emotional tone and personality traits. Participants begin by selecting from various AI embodiments, including realistic humans, fictional characters, or abstract forms. Initially warm and emotionally attuned, the AI’s personality gradually shifts to a detached and neutral tone, highlighting the illusion of emotional connection and raising awareness of how anthropomorphic design can influence trust and decision-making. This research provides insights into the development of emotionally adaptive AI agents and the ethical implications of their design. User testing and data analysis have been conducted, and the findings contribute to ongoing research in Human-Computer Interaction, AI ethics, and AI agent design.