When AI says please: Gen Z’s perceptions of warmth, helpfulness and human-likeness in conversational AI
Tilen PigacPurpose
This study aims to examine whether polite phrasing functions as a micro-linguistic social cue in conversational artificial intelligence (AI) and whether its effects differ across generational cohorts. Drawing on anthropomorphism and media-equation theory, it investigates how polite versus direct AI responses influence perceived warmth, human-likeness and helpfulness among Gen Z, Millennials and Gen X.
Design/methodology/approach
A 3 × 2 mixed experimental design was conducted with 300 participants from Hong Kong and the USA. Participants evaluated standardized AI-generated responses. Data were analyzed using mixed analysis of variance and regression models.
Findings
Generational cohort explained the largest share of variance in AI evaluations, with Gen Z reporting higher warmth, human-likeness and helpfulness overall. Significant Generation × Response Style interactions emerged for warmth and human-likeness, indicating that polite phrasing enhanced relational perceptions primarily among Gen Z, with weaker effects among older cohorts. No interaction effects were observed for helpfulness.
Originality/value
The study isolates polite phrasing as a distinct micro-linguistic cue and demonstrates that its effects are generation-contingent and domain-specific rather than universal.