Social Robotics Glossary

Android: A robot designed to look as much like a human as possible, including realistic skin, hair and facial features. Examples: Geminoid, Ameca, Nadine.
Anthropomorphism: Attributing human characteristics, emotions or intentions to non-human things, such as animals, objects or robots.
ASR (automatic speech recognition): Software that converts spoken words into text.
Backchannel: Small signals a listener gives while someone else is speaking, such as nodding, "mm-hm" or "right", to show they are following.
Cobot: A collaborative robot designed to work safely alongside people, usually in industry. Not necessarily social.
Contingency: A robot's behaviour responding directly and promptly to what a person does. Contingent responses are one of the strongest cues that make something seem alive.
Embodiment: Having a physical body that exists in the same space as the user. Also used for the idea that intelligence depends on a body interacting with the world.
End-of-turn detection: Predicting when a speaker has finished so the robot can respond without interrupting or leaving long silences.
Godspeed questionnaire: A widely used set of questionnaires for measuring how people perceive a robot, including anthropomorphism, animacy, likeability, perceived intelligence and perceived safety.
HRI (human-robot interaction): The research field that studies interactions between people and robots, and designs robots for them.
Humanoid: A robot with a body shaped broadly like a human, typically a head, torso, two arms and sometimes two legs, without necessarily looking realistic.
Joint attention: Two parties focusing on the same object or event, each aware the other is doing so. A foundation of human communication.
Legibility: How easily people can understand what a robot is doing and about to do from its movements and signals.
LLM (large language model): An AI model trained on huge amounts of text that can generate and understand language. Many social robots now use LLMs for open-ended conversation.
Media equation: The finding by Reeves and Nass that people treat computers and media as if they were real people and places, applying social rules automatically.
Multi-party interaction: A robot interacting with several people at once, which requires tracking who is speaking, who is being addressed and who is just listening.
Novelty effect: The initial excitement people feel with a new robot, which often fades. Short studies can overstate benefits because of it.
Proxemics: The study of how people use space and distance in social interaction, applied to how robots should approach and position themselves.
Socially assistive robotics (SAR): Robots that help people through social interaction rather than physical contact, for example coaching exercise or supporting therapy. Term introduced by Feil-Seifer and Matarić in 2005.
Social robot: A physical robot designed to interact and communicate with people following human social norms.
Theory of mind: The ability to understand that others have their own beliefs, desires and knowledge, which may differ from one's own.
TTS (text-to-speech): Software that converts text into spoken audio, giving a robot its voice.
Turing test: Alan Turing's 1950 proposal that a machine could be called intelligent if its conversation could not be told apart from a human's.
Turn-taking: The way participants in a conversation alternate between speaking and listening, coordinated by timing, gaze and intonation.
Uncanny valley: Masahiro Mori's 1970 idea that robots which look almost, but not quite, human cause a feeling of eeriness.
VAD (voice activity detection): Detecting when someone is speaking, as opposed to silence or background noise.
Wizard of Oz: A research method where a hidden human operator controls some of a robot's behaviour, while participants believe the robot is acting on its own.
Zoomorphic: Animal-like in form, such as the robot seal Paro or the robot dog aibo.

Back toThe guide →