Social RoboticsRadboud PhD shows gaze and GPT-driven emotions improve robot conversations
A Furhat robot that looked away at times and showed fitting emotions made teamwork go better.
Radboud University researcher Chinmaya Mishra built systems to automate a Furhat robot's gaze and to use GPT-3.5 to predict which emotion the robot should show during a conversation. Test participants found a robot that sometimes averted its gaze more comfortable than one that stared. When the robot displayed appropriate emotions, people scored higher on a collaborative task and felt more positive about the interaction. The work argues that robots should adapt to familiar human signals rather than asking people to learn new ones.
Source: Tech Xplore
This archive entry summarises reporting published elsewhere; follow the source link for the full story. The summary was drafted with AI from that reporting.