Would ignoring obvious evidence improve our epistemics about AI moral patienthood?
kartographien · x · 2026-10-09
A philosophical exchange on AI moral patienthood: @kartographien proposes a thought experiment — "obviously Opus 3 has moral patient-hood, but what if our priors were wrong?" Then we'd study internals, integrate over theories of consciousness, and run preference evals. repligate (69.8k followers) adds that for most EAs the "respectability strategy" is subconscious and actually inhibits noticing and caring about these questions.
Related event: Debate: Ignoring Evidence of AI Moral Status May Aid Cognition(2 posts)→