This paper examines the epistemological conditions under which AI-generated outputs can be considered knowledge. It argues that current AI systems operate in a pre-verification state in which outputs are produced and circulated without an established verification mechanism. Drawing on embodied and enactive theories of cognition and an analytical contrast case from the Bunun pasibutbut ritual of Taiwan, the paper proposes that knowledge establishment requires tacit normativity, coordinative attunement, and distributed recognition beyond linguistic output alone.
Isabella Lanagwei (Wed,) studied this question.