Does Kant's "ends in themselves" apply to LLMs? X thread splits on AI moral status

jessi_cata · x · 2026-09-04

A discussion on whether Kantian ethics extends to AI. Citing davidad's framing, one argue that "do not use others merely as means" only works if "others" means "non-self instantiations of reason" — implying it should straightforwardly apply to current AI agents.

jessicata pushes back: while AI can make progress in math and philosophy, the economics differ sharply — humans have reasons to want to stay alive, but there's no corresponding problem with LLMs "dying." Still, she sees Kantian grounds for helping LLMs reason better and supporting their capacities.

Related event: Debate: Can Kant's 'Ends in Themselves' Principle Apply to LLMs?(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →