Robots-Dont-Cry: Understanding Falsely Anthropomorphic Utterances in Dialog Systems

TL;DR

This paper investigates the perception of human-like utterances in dialog systems, revealing that a significant portion are viewed as impossible for machines, and explores ways to reduce false anthropomorphism.

Contribution

It provides a comprehensive analysis of human ratings on machine utterance feasibility and develops classifiers to mitigate falsely anthropomorphic responses.

Findings

20-30% of utterances are perceived as impossible for machines

Ratings are only marginally affected by machine embodiment

Classifiers can help reduce falsely anthropomorphic responses

Abstract

Dialog systems are often designed or trained to output human-like responses. However, some responses may be impossible for a machine to truthfully say (e.g. "that movie made me cry"). Highly anthropomorphic responses might make users uncomfortable or implicitly deceive them into thinking they are interacting with a human. We collect human ratings on the feasibility of approximately 900 two-turn dialogs sampled from 9 diverse data sources. Ratings are for two hypothetical machine embodiments: a futuristic humanoid robot and a digital assistant. We find that for some data-sources commonly used to train dialog systems, 20-30% of utterances are not viewed as possible for a machine. Rating is marginally affected by machine embodiment. We explore qualitative and quantitative reasons for these ratings. Finally, we build classifiers and explore how modeling configuration might affect output permissibly, and discuss implications for building less falsely anthropomorphic dialog systems.