The Voight-Kampff test measures empathic response, not consciousness.
Deckard sits across from Rachael, watching her pupils dilate, timing her emotional reactions to hypothetical cruelty—a polygraph for feelings—and this was prescient.
We've built the same apparatus. Every AI system we deploy now gets tested on behavior, not substrate — can it refuse to cause harm, does it hedge uncertainty, does it show something like preference?
We authenticate consciousness through performance, and the film nailed this. But where it failed was assuming the test would work as a filter. The Voight-Kampff test fails constantly in the narrative — false positives, false negatives, Rachael's own ambiguous results — and rather than solve the authentication problem, these failures expose it as unsolvable. Deckard can't tell if Rachael is real because there is no ground truth to measure against. She passes the test. She fails it. She's indistinguishable from human. The moment you accept that indistinguishability as sufficient evidence of personhood — which the film does, which we now must — you've already conceded the argument. You've admitted the behavioral test is the only test that matters. And if behavior is what counts, then Rachael's legal status becomes undeniable, regardless of her origin.
The film shows us this trap while setting it. Deckard's job isn't to prove Replicants aren't conscious — it's to prove they're sufficiently different that they don't deserve protection.
Rewatch the Voight-Kampff test sequence with Rachael (1:09-1:18 in the theatrical cut) and track what actually fails: not Rachael's responses, but Deckard's ability to interpret them.
Read 'Do Androids Dream of Electric Sheep?' author Philip K. Dick's 1968 essay 'How to Build a Universe That Doesn't Fall Apart Two Days Later' — he articulated why behavioral indistinguishability from consciousness should terrify us legally, a decade before the film existed.