Toward a Neurocognitive Account of CASA: EEG Evidence from Human and ChatGPT-Enabled Social Robot Conversation
Drawing on the Computers Are Social Actors (CASA) paradigm, this study compared electroencephalographic (EEG) responses during naturalistic conversation with a human interlocutor and Pepper 1.8a, a social robot powered by a customized ChatGPT API. Female university students in South Korea discussed merit- versus need-based scholarships with both interlocutors in a within-subjects experiment. Human interaction elicited higher frontal theta and higher central-parietal alpha power than robot interaction, which may indicate greater demands on interaction monitoring and adaptive control, as well as a different attentional-regulatory state, during human interaction. Frontal alpha asymmetry did not significantly differ between conditions, providing no evidence of a difference in approach-related motivational orientation. In exploratory analyses, greater frontal alpha asymmetry during human relative to robot interaction was associated with a smaller difference in enjoyment. These findings provide downstream neurophysiological evidence relevant to CASA: interaction with a social robot was associated with EEG responses linked to ongoing interaction-related processes.
Authors
- Sungbin Youk (ORCID: https://orcid.org/0000-0001-8859-2757)
- Sun Kyong Lee (ORCID: https://orcid.org/0000-0002-8645-8922)
- Olivier Jiyoun Jung (ORCID: https://orcid.org/0009-0003-6532-5068)
- Yeowon Jie (ORCID: https://orcid.org/0009-0009-2322-8706)
Institutions
- Ewha Womans University (KR)
- Korea University (KR)
Publication Details
- Journal
- International Journal of Human-Computer Interaction
- Published
- 2026-11-19
- DOI
- https://doi.org/10.1080/10447318.2026.2762653
- Primary Topic
- Social Robot Interaction and HRI
- Type
- article
- Field-Weighted Citation Impact
- 0.00