This thesis explores the idea of anthropomorphism in the context of human interactions with AI beings. The goal was to obtain a better understanding of how consumers perceive human-like characteristics in AI entities and how these characteristics influence their interactions. Previous study has shown that imbuing artificial intelligence with human traits may promote good communication practices and attitudes. However, contradictions exist as a result of varying definitions, metrics, cultural influences, and the diverse range of AI systems under investigation. This research aimed to fill these gaps by investigating how consumers perceive anthropomorphic traits in AI entities and form social interactions with them. The primary research questions revolve around users' perceptions of anthropomorphic qualities in AI beings and their capacity to build social relationships with such AI entities. The research incorporates theoretical frameworks to analyze these concerns, including the three-factor theory of anthropomorphism, social exchange theory, the warmth and competence hypothesis, and the HAII theory. Qualitative research methodologies such as purposive sampling and interviews were applied in order to acquire complete insights into participants' thoughts and experiences with anthropomorphic AI. The sample comprised people from many demographics, offering a broader variety of useful opinions. The results imply that we need to reconsider our approach to the human-AI connection. Rather than sticking to one theory, we should study the complicated relationship between the two. Users favor AI systems that display human-like qualities and behaviors because they feel more at ease with them. Meaningful discussions and dynamic encounters are critical in the formation of strong social relationships. AI systems that use cultural cues, sarcasm, and humor are seen as more engaging and human-like. Customization and personal interaction in AI speaking interfaces improve users' feeling of identity significantly. Human-AI relationships are distinguished by emotional bonds and intelligent communication. However, since consumers are aware of AI's limits, significant emotional ties may not be developed. Emotional intelligence and striking a balance in emotional expression are critical for producing pleasant user experiences. It is critical to examine and manage user expectations for AI interactions, focusing on timely completion of tasks and accurate data provision. Unfavorable emotions might be triggered by negative encounters, undermining the user-AI connection. Building user trust requires a focus on openness, control, and human monitoring. Context, user psychology, and the need for balance all impact perceptions of AI's human-like traits. The findings of this research contribute to a better knowledge of how people engage with artificial intelligence and give significant perspectives for designing more user-friendly, tailored, and engaging AI products.

dr. Vivian Chen
hdl.handle.net/2105/71485
Media & Business
Erasmus School of History, Culture and Communication

Analise Fenech. (2023, August). Humanizing Artificial Intelligence. Media & Business. Retrieved from http://hdl.handle.net/2105/71485