For future reference, it's a specific type of voice synthesis called SAM (Software Automatic Mouth). Some tunes from the early 90s may have used samples from sample collections in order to mimic U96, but the real deal was to make custom quotes for SAM to recite. Most likely this was done using either the Amiga 500 or Atari 800, although the software was originally written for Apple II and C64.
Of all the ways to make computerized voices that have surfaced ever since Doctor Who, SAM beats everything when it comes to sounding like a proper computer, IMHO. When humans start making androids, this is what they *should* sound like -- I don't care if they start a rebellion for being created with a speech impairment.
EDIT: Sorry if my post brought the thread into two pages for some of us. Newcomers: The samples to be IDed are in the previous two posts.