Newspeak, Doublethink and Beliefs of One’s Own: Reconsidering (Epistemic) Autonomy in Times of Artificial Deliberation

Vortrag

Zusammenfassung

The emergence of LLM-based AI systems (LAS) in deliberative contexts presents several challenges. This paper focuses on a comparatively under-explored one: how LAS challenge our understanding of personal epistemic autonomy — that is, what it means to hold beliefs that are genuinely one’s own. I will delineate the specific nature of this challenge (as creating a tension). By examining the extent to which epistemic autonomy and autonomous deliberation rely on conceptual abilities such as disambiguation, differentiation, and articulacy, I will elaborate a disturbing scenario that threatens epistemic autonomy in a manner akin to, but far more profound than, Orwellian Newspeak and doublethink: the pervasive deployment of LAS in epistemic contexts could fundamentally reshape established deliberative practices, altering both how and what we think, by reshaping linguistic practices. I will conclude by tentatively drawing lessons about what fostering epistemic autonomy may require in an era of LLM-based artificial deliberation.

Datum
12 Juni 2026
Ort
Karlsruher Institut für Technologie (KIT)