THE COLLECTIVE SOCIAL BRAIN AND THE EVOLUTION OF POLITICAL POLARIZATION

Abstract

Purpose of the study: In this exploratory research paper, we utilize the capabilities of ChatGPT-4, an advanced artificial intelligence model, to investigate the collective social brain hypothesis in the context of political polarization. We posit that human groups can be broadly categorized into two response profiles that correspond to two halves of a “collective social brain”, one half of which uses a problem-solving method (system I thinking) that tends to use consensus for evaluating truth in areas in which they feel vulnerable and in need of protection. The other half tends to use system II thinking to think independently in those same areas as they don't feel vulnerable. These problem-solving methods simply come to different conclusions given the same information. Both thinking types are useful for solving different problems, but are harmful when applied to the wrong problems. Groups at the size of the ancestral tribes we evolved in can switch to whatever thinking System Is optimal, but at the size of current societies these switching mechanisms break down and exchanging more information (news, social media, etc.) just leads to more polarization. Methodology: Leveraging AI-based simulations, we collect and analyze data from social media discourse, categorizing responses into these response profiles. Main Findings: Our simulated findings reveal distinct response profiles prevalent in comments, varying by topic, platform, geographical location, and time of posting. We observe a significant association between the type of reasoning used and the topic of the post. Our research supports the collective social brain hypothesis and highlights the potential for mitigating polarization through the recognition and accommodation of differing reasoning styles. Research limitations