The double-edged sword of artificial intelligence: How chatbots are changing political consciousness in the age of identity politics

Research shows that AI-based chatbots can change political attitudes through private, neutral conversation, bypassing identity-based defense mechanisms. However, concerns arise about algorithmic biases and hidden manipulations, as evidenced by studies on partisan bias in models and the phenomenon of 'sectoral lock-in'.

New research shows that conversations with AI-based chatbots can change political attitudes, even in an era of identity politics where opinions are perceived as part of personal identity. A study published in December 2025 in the journal Nature, led by Chiara Varrio and Alessandro Nai, found that the chatbot succeeds in persuading by presenting facts and rational arguments, without emotional manipulation, thanks to the private and neutral space it creates. However, a 2025 study from the University of Washington showed that partisan-biased models can influence users to adopt views contrary to their original identity. A simulation by ChatOptic in August 2026 revealed a phenomenon of 'sectoral lock-in', where models present to sectoral personas only parties that directly appeal to their sector, ignoring others. The article warns of algorithmic biases and hidden manipulations, especially given the growth of the field of advocacy in the age of AI (GEO).

The double-edged sword of artificial intelligence: How chatbots are changing political consciousness in the age of identity politics