This research investigates how artificial intelligence interprets and expresses the concept of humanity through visual art, and how people, in turn, perceive the humanity of AI based on these outputs. The project begins with the collection of approximately two thousand survey responses, in both Chinese and English, in which participants shared their personal views on the meaning of humanity. These responses were aggregated and used as input for an AI image generation process, resulting in a series of visual artworks intended to embody collective human perspectives.

To examine audience perceptions of these AI-generated works, semi-structured interviews were conducted with master-level design students who regularly engage with creative technologies. Participants were asked to interpret the artworks without prior knowledge that the images were created by AI, ensuring that responses were guided by the visual content itself rather than by preconceived notions of machine authorship.

The qualitative data were analyzed using Reflexive Thematic Analysis (RTA), through which recurring interpretive patterns were identified and grouped into six themes. To complement this analysis, participants also completed the Godspeed questionnaire, providing quantitative measures of anthropomorphism, animacy, and perceived intelligence. By correlating thematic interpretations with these perception scores, the study highlights how specific visual features—such as symbolic richness, emotional resonance, relational elements, or mechanistic qualities—influence the extent to which people attribute humanity to AI.

Findings suggest that judgments of AI’s humanity are shaped less by technical fidelity than by the capacity of visual features to evoke emotions, cultural references, and symbolic associations. Participants who linked images to personal or collective human experiences tended to ascribe greater humanity to the AI, while those who focused on surface resemblance or structural flaws were more skeptical. This indicates that perceptions of AI’s humanity depend on interpretive depth rather than superficial resemblance.

The project contributes to human–computer interaction research by demonstrating how qualitative interpretations of AI-generated art can reveal underlying frameworks through which people negotiate the boundary between human and machine creativity. It also provides practical insights for the design of AI systems in artistic and communicative domains, suggesting that fostering symbolic and emotional resonance may be more impactful than pursuing purely technical realism. Ultimately, this study positions AI art as a valuable lens for exploring evolving understandings of humanity and for guiding more responsible and human-centered development of creative AI systems