The Alien Intelligence Thesis

Lorraine Daston has argued that the fundamental error in evaluating artificial intelligence is the attempt to assess it using the frameworks built for human knowledge production—and that the error persists because those frameworks are so deeply embedded that they have become invisible. The evaluative apparatus of human epistemic culture, from peer review to professional reputation to the rules of scholarly citation, assumes a knowledge producer who is a social agent: embedded in relationships of trust and accountability, capable of being held responsible for errors, subject to the norms and sanctions of a community, and possessed of the social and personal stakes that motivate care. AI is none of these things. It has no institutional affiliations, no reputation at stake, no colleagues who can be disappointed or impressed, no professional identity that error threatens. It produces its outputs through a process that is not social, not iterative in the relevant sense,