The article starts by saying that Apple's speech synthesis was innovative when it came out in 1993... but the original Mac spoke when it was first introduced in 1984:
Secondly that "My voice is my password" protection in OS 8.x was a joke. My friends and I used to fake it out all the time just by mimicking each other's voices. I don't see why a recording wouldn't get by it either.
Lastly, the state of speech synthesis has come a LONG way since OS 7, yet Apple's voices haven't seemed to improve all that much. The samples at AT&T Labs bear this out. I'd really like to see them step it up a nothc and improve the "Macintalk Pro" voices a LOT, or just license the AT&T Labs work for better speech synthesis.