Would an intelligent robot deliberately fail the Turing test?

LMAO! I've wondered about a more fundamental question. If a human is unable to determine with greater than 50% certainty whether a computer interviewee is human or not, did the computer pass the Turing test or did the human fail it?

The obvious answer to your question is yes, under the assumption that some device could be built that exhibited human intelligence--complete with self-awareness and human-style emotions. Since nobody knows exactly how those work in humans, I think that's an unfillable prescription.

"Could" is probably more likely then "would". The correct answer (as to whether an intelligent robot "would" fail a Turing test) would depend on the intelligent robot's motivation for allowing itself to be subjected to the Turing test. If said robot were trying to trick a person into thinking that it isn't semi-sentient or fully sentient, then it would naturally surmise that failing the Turing test is an easy way to accomplish the task (or at least a step in the right direction).