Post History

I would find this approach highly unrealistic. For one, those reel-to-reel tape drives do not EVER record sound; they are strictly digital. I worked with them extensively in the 1970's; for a time I was one of many people using computers to predict weather and compute flight plans. Speech synthesis for computers has been around almost since they began, in the 1950's. Below is a link. Back in the 70's, we already had digital microphones that could record individual words and computer systems that could play them back to speakers in whatever combinations desired, for example to read a telephone number out loud, or spell out a word. 

If you wish to show the limitations of 1970's technology, I would suggest the AI have a very limited vocabulary of a few hundred words, like a child. Storage was always the issue.

> Noriko Umeda et al. developed the first general English text-to-speech system in 1968, at the Electrotechnical Laboratory in Japan. In 1961, physicist John Larry Kelly, Jr and his colleague Louis Gerstman used an IBM 704 computer to synthesize speech, an event among the most prominent in the history of Bell Labs. https://en.wikipedia.org/wiki/Speech_synthesis