The Voight-Kampff test measures pupil dilation, blink response, and involuntary facial micro-expressions triggered by questions about animal suffering.

It's a humanism test built on the premise that feeling is a biological signature you can't fake — yet Rachael passes by doing exactly that.

She fakes the response. And the film never cleanly resolves whether she's a replicant performing humanity or a human with implanted memories — because the distinction stops mattering the moment the performance holds up to scrutiny.

Scott's actual insight sits in the margins of the plot: he designed a system that can't distinguish between authentic empathy and its perfect simulacrum, then populated his world with creatures who pass it anyway. The Voight-Kampff doesn't solve the alignment problem — it proves there is no alignment problem to solve, because alignment isn't about internal states — it's about behavioral coherence.

We're building that test right now. Every chatbot trained to express concern about your wellbeing, every recommendation algorithm designed to seem to understand your taste, every AI system engineered to perform emotional intelligence — we've stopped asking if they actually care, and we stopped the moment the performance became indistinguishable from caring.