Mistaking minds and machines: How speech affects dehumanization and anthropomorphism

Area of Science:

• Cognitive Psychology
• Human-Computer Interaction
• Linguistics

Background:

• Dehumanization involves treating human minds like machines.
• Anthropomorphism involves attributing humanlike minds to machines.
• Paralinguistic cues in speech are central to perceived mental experience.

Purpose of the Study:

• To investigate how humanlike voice cues influence perceptions of human vs. machine creators.
• To determine if voice, compared to text or visual cues, is critical for inferring personhood.
• To explore the impact of speech's naturalistic paralinguistic features on these inferences.

Main Methods:

• Four experiments were conducted comparing perceptions of creators based on voice, text, and video stimuli.
• Participants evaluated whether the creator of a message was human or a machine.
• The presence or absence of voice, and specific paralinguistic cues, were manipulated.

Main Results:

• Hearing thoughts expressed in a human voice increased the likelihood of inferring a human creator compared to reading text.
• Visual cues in video clips did not enhance human inference over text alone.
• Removing naturalistic paralinguistic cues from speech eliminated its humanizing effect.

Conclusions:

• Speech, particularly its paralinguistic elements, plays a crucial role in attributing humanlike minds.
• Text-based communication risks dehumanization by obscuring human mental experience.
• Speech-based machine interactions risk anthropomorphism due to the perceived presence of a humanlike mind.