Testing Conformity in AI-Integrated Political Group Chats

This experiment investigates whether participants adjust their political expressions when placed in group chats with either liberal or conservative-dominated opinions. The aim is to examine whether closed online settings, such as those simulating WhatsApp or Telegram conversations among friends or colleagues, induce self-censorship or conformity, thus amplifying the perceived polarisation of society. Political shift will be analysed using GPT-4-based semantic analysis.