Counter-closure principles in the age of complex software systems: a generalized challenge from AI

The rapid advancement of artificial intelligence has brought a host of new epistemological challenges. One particularly pressing question is whether, and to what extent, AI systems can serve as sources of epistemic goods. Can they effectively transmit knowledge or understanding? And if they do not possess these epistemic goods themselves, can they still generate them for human users? This article explores these questions by critically examining the constraints posed by counter-closure principles – epistemological principles that allegedly cast doubt on the epistemic potential of AI. By addressing these principles, we aim to lay the groundwork for a systematic inquiry into the social epistemology of AI.