The Epistemic Politics of AI Anthropomorphism
The episode discusses Donna Bye and Levin Kuhlmann's paper, "The Epistemic Politics of AI Anthropomorphism." The hosts explore how institutions use 'epistemic authority' to decide what is a healthy way to interact with AI. They detail a self-validating loop where framing discourages engagement, and propose five commitments for design changes, including weighing the costs of under-ascription versus over-ascription.