Introduction. GenAI is increasingly embedded in daily life, yet its ostensibly neutral design often becomes a site for gendered interpretations. This study examines how Gen Z users in China perceive and construct gendered traits in GenAI, situating their accounts within broader concerns of neutrality and equity in information society. Method. Semi-structured interviews were conducted with 12 participants who regularly use multiple GenAI platforms. Discussions explored textual, visual, and auditory cues, user projections, and contextual triggers. Analysis. Thematic analysis, informed by Butler’s theory of performativity and Haraway’s cyborg imaginary, was applied to 73,480 words of transcripts. Coding identified patterns in participants’ accounts of gendered cues, user projection, contextual attribution, and reflections on neutrality. Results. Participants consistently mapped rational, didactic tones onto masculine authority and empathetic language onto feminine care. Avatars and voices anchored gender perceptions, while mimicry highlighted both adaptability and artificiality. Attribution also drew on cultural repertoires, personal experiences, and task contexts. Conclusion. Neutrality in GenAI does not erase gender but becomes a site of projection and negotiation. The findings show how such systems can reproduce stereotypes while also enabling more inclusive, post-binary imaginaries, underscoring the need for critical AI literacy and designs that advance information equity.