HMC in Context

Gendered and ethnic-racial divisions of human domestic labor influencing social responses to voice-based agents: Inconsistency effects and prejudice intensification

Ample research has supported and challenged how users apply human-human communication scripts to their interactions with machines. Still, few have examined social responses to machines in specific contexts - the places and purposes of machine use and their socio-historical backgrounds. This project employs the Media Are Social Actors paradigm to obtain experimental evidence for the effects of machine-signaled international social identities (i.e., gender and ethnic-racial identities) in the context of inequalities (i.e., domestic labor).

​

HMC over Time

Children’s parasocial relationships with voice-based agents:

A longitudinal experimental test of relationship development, prejudice reduction, and learning improvement

Theories and empirical findings have suggested how the frequency and dynamic of HMC may change over time as users accumulate more experience with interactive media technologies.​ This transformation may be particularly evident among children, who experience a drastic increase in technology knowledge and exposure from the minimal. This project investigates how children develop parasocial relationships with machines over time and how machine-signaled ethnic-racial identities might alter this process.

​

HMC - the Human-Machine Distinction

​AI-generated realistic images: Human detection and effects on persuasion across cultures

The functional and perceptional distinctions between humans and machines have become increasingly nuanced. Users' beliefs in machine heuristics (e.g., machines are accurate) vary with user technology experience and the nature of machine tasks. This project examines how users perceive the source (i.e., human or machine) of realistic images, which entail both realism and creativity, and how these source perceptions influence image persuasiveness in the context of health communication.

​​

HMC - the Effects of Multiple Social Cues

How should AI talk? Examining disclaimers and anthropomorphic design in shaping trust toward AI chatbots

Today's machines have larger bandwidths and, consequently, can display more social cues than their predecessors. This study examines how the quality and quantity of multiple social cues may influence user trust in AI chatbots. Results revealed a significant interaction between language patterns and anthropomorphic identity markers: for a chatbot that displayed nonsocial disclaimers, high anthropomorphic identity markers indirectly enhanced user trust by increasing the chatbot's social presence.