In the experiments, we compared two types of robots, human-like robots and machine-like robots. We designed this prototype from 4 aspects.
\subsection{Appearance}
The user's first impression of the robot is its appearance. Therefore, the appearance of the robot is crucial in helping the user to develop the correct mental model with the robot. In this study, a Pepper robot \cite{softbank_pepper_2022} was employed, the machine-like type robot maintains its original appearance, whereas the human-like type robot dresses in clothing with noticeable "human" traits, such as a tie and a hat.
\subsection{Behaviour}
The user's perception of the robot's identity during the interaction is heavily influenced by the robot's behaviour, such as the robot's body movements when talking to the user. Specifically, the robot's gaze and gestures have an impact \cite{kompatsiari_importance_2017, salem_friendly_2011}. Likewise, \cite{fischer_why_2019} suggested that collaborative robots should show social cues, which may involve the display of emotions. Therefore, in this study, the majority of behaviours are kept the same for both types of robots. The only different behaviour is the dance of the robot. Human-like robot dance has the following characteristics: human-quality motion, flowing, organic, natural, and curved lines, whereas the machine-like robot dance has: precision, control, proximity and safety \cite{abe_pilot_2020}.
\subsection{Voice}
Hearing is one of the most important human senses, and a robot's voice can also affect the user's perception of its identity. In this study, the voice was adjusted by the built-in voice settings. The \textit{naoenu} setting is a distinctly human voice with tone and emotion which is used for the human-like robot. The \textit{naomnc} setting is apparently a synthesized voice which was typically considered to be machine-like \cite{walters_human_2008} is employed by the machine-like robot.
\subsection{Dialogue}
Fischer \cite{fischer_what_2006} has summarised the literature on the distinctions between human-to-human communication (HHC) and human-computer interaction (HCI) in several aspects. In this study, the dialogue with a machine-like robot with the following features compared to the human-like robot: less polite, with fewer words per conversation, a smaller lexicon, fewer syntactic, more simplification, including a little over-specification, and technical disclosure rather than emotional disclosure. Table \ref{edialogue} shows an example segment of the dialogue. It can be noticed that this machine-like robot has a bit of over-specification and considerable technical disclosure.
