Since the beginning of the chatbot hype, there has been a lively discussion about how “human-like”
chatbots should be. To make the interaction more natural and enhance user experience (UX), many
chatbots are given a name, are represented by a human-looking avatar, or are endowed with a
personality in general. However, these human-like or anthropomorphic cues can lead to unrealistic
user expectations that often turn into frustration and annoyance when the chatbot fails. Therefore,

designers, developers, and organizations are often unsure about the “appropriate” level of human-
likeness of their chatbot and its potential effects on user interaction and behavior. To address this

challenge, this talk will present theoretical and practical insights from several years of research on the
anthropomorphic design of chatbots at the Karlsruhe Institute of Technology (KIT). First, the talk will
provide an overview of the many different human-like or anthropomorphic cues in the design of
chatbots. Second, the underlying psychological mechanisms that lead users to (subconsciously)
respond to these cues when interacting with a chatbot will be explained. Finally, key findings from
several lab and field studies with real chatbots on the impact of these cues on user interaction and
behavior will be presented. This talk will be interesting for everyone who wants to learn more about
the theoretical aspects of human-chatbot interaction or aims to convince users to engage with their
existing chatbot.