You On AI Field Guide · The AI Consciousness Test The You On AI Field Guide Home
TxtLowMedHigh
CONCEPT

The AI Consciousness Test

Susan Schneider and Edwin Turner’s proposal to test machine consciousness by isolating a system from human consciousness-talk during training and then checking whether it independently generates the concepts that naturally arise from having an inner life.
The AI Consciousness Test, or ACT, is philosopher Susan Schneider’s most ingenious response to the epistemic wall that the hard problem of consciousness erects between behavior and experience. The problem it addresses is structural: any system trained on human text has absorbed the entire reservoir of human consciousness-talk, and can therefore produce fluent, nuanced descriptions of inner life without having any. Behavioral evidence for consciousness is precisely worthless in a system optimized to produce the behavioral signatures of consciousness. The ACT sidesteps the imitation problem by changing the epistemic situation: an artificial system is deliberately isolated from human discourse about consciousness, souls, survival, and out-of-body experience during training, and then asked the kinds of questions that arise naturally from having an inner life. If such a “boxed-in” system independently conceives of the possibility that it might persist beyond the destruction of its hardware, that there could be a difference between itself and a perfect copy, that identity might
← Home0%
CONCEPTBook →

Keep reading with YOU ON AI

Unlock the full book, 10,000+ field-guide entries, and a 1000+ thinker library. If you have a book code, register now — it takes a minute.

Register with book code Sign in