Speaker
Description
The rapid adoption of generative AI fundamentally changes how we search for and evaluate information. As conversational AI systems replace traditional search engines, the need for independent source analysis diminishes, raising critical questions about epistemic authority and trust. This presentation analyzes generative AI as a new epistemic authority. Drawing on Miranda Fricker’s framework of epistemic injustice (Fricker, 2017), it explores how trust in AI is constructed. Using the concept of "testimonial sensibility," it argues that conversational interfaces trigger spontaneous, unreflective judgments of algorithmic credibility rather than critical inference. Additionally, empirical research shows that the precise format of AI-generated answers acts as a powerful cognitive anchor, artificially inflating user trust. Discussing the shift toward AI as knowledge brokers (Perreault et al., 2025), the presentation highlights the risks of pre-emptive testimonial injustice, the concentration of epistemic power, and threats to democratic information access. Ultimately, as AI is increasingly perceived as a potential "knower," combating these emerging epistemic imbalances demands both individual critical self-reflection and the implementation of robust structural mechanisms.
| Presenting author | Maciej Krzemiński |
|---|
