The HMIHY system falls somewhere between these two extremes.
Hence, the goal of the HMIHY user study was to come up with an initial greeting that lets callers know they are talking to a machine, not to a human.
The HMIHY simulation tested two ways of communicating to callers that they were speaking to a machine: One said explicitly it was a machine in the first announcement heard by the caller; the other used computer-generated text-to-speech output to make the system "sound" like a machine.
Still unknown is how callers might have responded to a text-to-speech version of the simple HMIHY prompt.
A subsequent study of HMIHY interfaces was conducted at AT&T Labs in 1997 to further refine the wording of the system's initial greeting and test another alternative: playing an "audio logo"(2) or sound effect at the beginning of the greeting as a way of cuing callers they are communicating with an automated system.