aceModerate evidence
More
Android with facial mimicry made chats with ChatGPT feel more positive, an experiment finds
Study title: Android embodiment with facial mimicry may enhance emotional responses during conversations with AI agents
- Evidence type
- Controlled or validated comparison
- Sources
- 1 sourcePeer-reviewed
- Published
- 2 min read
- Truvace record
- TRV-2026-2656cite or check it
Study title: Android embodiment with facial mimicry may enhance emotional responses during conversations with AI agents
Researchers tested whether a physical face changes how people feel when they talk with a conversational AI. Participants held positive emotional conversations with ChatGPT, delivered either through an android robot that mimicked facial expressions or through a blank display.
Read the study abstract
The potential of artificial intelligence (AI) agents to engage in emotional conversations with humans is attracting increasing attention. Previous studies suggested that conversational agents powered by large language models such as ChatGPT can engage in emotionally competent conversations with humans. However, whether verbal communication alone is sufficient or whether nonverbal communication might exert additive effects on emotional conversations remains unknown. We conducted an experiment to examine whether a…
The gain
In an experiment on positive conversations with ChatGPT, participants who talked to it through an android robot with facial mimicry rated their feelings as more positive and showed more smiling-muscle activity than those using a blank display.
In brief
Participants in the android condition reported significantly more positive valence ratings and showed greater activity in the zygomaticus major, a cheek muscle involved in smiling, than those facing the display.
The authors describe the results as initial evidence. Arousal, frown-muscle activity and skin conductance did not differ significantly between the two interfaces, so the effect was limited to positive feeling and smiling.
Main points
- Participants had positive emotional conversations with ChatGPT through either an android robot showing facial mimicry or a blank display.
- Self-reported valence was significantly more positive with the android than with the display.
- Zygomaticus major EMG activity, linked to smiling, was greater during conversations with the android.
- Arousal ratings, corrugator supercilii EMG activity and skin conductance level showed no significant differences.
- The authors call the findings initial evidence that android embodiment may enhance positive emotional responses.
The full story
The study addresses an open question: earlier work suggested that large language model agents such as ChatGPT can hold emotionally competent conversations, but it was unknown whether words alone suffice or whether nonverbal cues add to the effect.
To test this, the team compared two interfaces for the same ChatGPT conversations. They measured emotional experience through valence and arousal ratings, and body responses through facial electromyography of the corrugator supercilii and zygomaticus major muscles plus skin conductance level.
Only the valence and zygomaticus measures differed significantly, both favouring the android. Arousal, corrugator activity and skin conductance did not, which leaves the evidence for a broader emotional effect unsettled.
Sources
1 on file- Telematics and Informatics Reportsdoi:10.1016/j.teler.2026.100379
Talk it over
Who gets this benefit, and who was left out of the study or story?
Android with facial mimicry made chats with ChatGPT feel more positive, an experiment findshttps://truvace.com/item/android-embodiment-with-facial-mimicry-may-enhance-emotional-responses-during-co
Who gets this benefit, and who was left out of the study or story?
The gain In an experiment on positive conversations with ChatGPT, participants who talked to it through an android robot with facial mimicry rated their feelings as more positive and showed more smiling-muscle activity than those using a blank display.
Answer in the threadHow do we know? What can this kind of evidence show, and what can’t it?
The evidence This rests on a controlled comparison, from Telematics and Informatics Reports.
Answer in the threadWhat does it leave open? What else would you want to know before acting on it?
The record’s own caveat The authors describe the findings as initial evidence, and arousal, corrugator EMG activity and skin conductance level showed no significant effects.
Answer in the threadHave you seen AI used this way where you study, work or live? Did it go the same way?
Answer in the thread
The questions are the same on every record. The details under them are taken from this record as written.
What readers say