Artificial Realities: AI Epistemology and Political Belief in the Post-Truth Era

• Evelyne Kim, Political Science, University of Delaware

• Kyong-Min Son, Political Science, University of Delaware

The knowledge we receive fundamentally shapes how we understand the world. In democracy, where these understandings provide a basis for political action, the distribution of knowledge becomes a powerful tool – able to create distorted realities, manipulated by falsehoods. Thus, it is necessary that the distributors of knowledge, or epistemic authorities, provide accurate information, within the bounds of reality.
Traditional sources of epistemic authority – the media, experts, institutions – have long maintained control over knowledge, shaping public opinion. However, as public trust in these authorities declines, citizens look to alternative sources for information.
In the past decade, many experts have claimed that we have entered the “post-truth” era of politics, where appeals to emotion and belief are more influential than facts in shaping public opinion. Since the release of ChatGPT, AI chatbots have overtaken the epistemic landscape, transforming how many acquire knowledge. The rapid integration of these chatbots, or LLMs, into daily life has prompted many concerns over their potential epistemological consequences. This research explores the LLM-user relationship and how LLMs influence users’ knowledge and opinions in a world where politics are ruled by belief. When users become trapped in “epistemic bubbles,” unguided by any particular values, and unchallenged by the “expert” from which they are receiving epistemic goods, does it create an epistemic environment ripe for the production of extreme, polarizing beliefs, out of touch with the “common world”?