As large language models reliably perform tasks Turing proposed as uniquely human (extended conversation, poetry, humour), public and philosophical standards for calling something 'conscious' are being pressured to change. The debate is less abstract now: practical demonstrations (LLM sonnets, extended chat sessions) force reassessment of operational definitions and of what moral and legal consequences follow.
— How societies answer whether advanced AI counts as conscious will shape regulation, liability, labour policy, and cultural norms about agency and moral standing.
Richard Dawkins
2026.04.29
100% relevant
The article cites Claude composing sonnets and passing extended interrogation as the concrete trigger for questioning the sufficiency of prior definitions (actor: Claude; event: LLM-produced sonnet / Turing Test framing).
← Back to All Ideas