“I know you're not human”: Non-knowledge, enlightenment, and the limits of social AI

This paper examines social AI through the lens of non-knowledge, arguing that current Responsible AI discourse overemphasises knowledge, control, and transparency while overlooking the epistemic uncertainty that structures human–AI interaction. Drawing on Enlightenment epistemology (Locke, Hume, Smith) and Luhmann’s theory of nonknowledge, it analyses Replika and Kindroid as distinct models of relationality. The paper shows that social AI does not simulate relationships, but co-produces them through interaction under conditions of epistemic indeterminacy. Responsible AI, therefore, emerges not as the elimination of uncertainty, but as ethical action within it.