Inter-brain Data of Human–AI Social Group Polarization

This study employed functional near-infrared spectroscopy (fNIRS) hyperscanning to investigate the behavioral characteristics of group polarization and the underlying inter-brain neural synchronization during group discussions in human–AI societies.Research ObjectivesThe objective of this study was to investigate the neural mechanisms underlying group polarization in human–AI societies by examining changes in inter-brain synchronization (IBS) during group discussions and its relationship with the degree of group polarization.

The findings provide neuroscientific evidence for understanding how AI participation in social interactions influences human group decision-making and attitude formation.Data Collection

Methods

A laboratory-based experimental design was adopted in this study. Participants first completed an initial attitude assessment on controversial social issues. They were then assigned to four-member discussion groups consisting of two human participants and two AI agents.

The groups discussed target topics through an online chat platform. During the discussions, cerebral hemodynamic signals from the two human participants were simultaneously recorded using a Shimadzu LABNIRS functional near-infrared spectroscopy (fNIRS) system. After the discussion, participants completed a post-discussion attitude assessment, and group polarization indices were subsequently calculated.ParticipantsA total of 37 groups (74 healthy university students) were included in the final analysis.

All participants were right-handed, had normal or corrected-to-normal vision, and reported no history of neurological or psychiatric disorders. Written informed consent was obtained from all participants prior to the experiment. The study was approved by the Institutional Review Board (IRB) or Ethics Committee of the authors' institution.Dataset ContentsThis dataset contains behavioral data, attitude rating data, group polarization indices, and fNIRS hyperscanning data collected during the experiment.

The dataset can be used to investigate group polarization in human–AI societies, inter-brain synchronization, and their underlying neural mechanisms. It also provides valuable resources for methodological development, data analysis, and replication studies in related research fields.