I'm Alan Alder, and this is clear and vivid conversations about connecting and communicating.

They create this machine called the voter and demonstrated at the 1939 World's Fair.

And it kind of looks like a little organ.

There were two dozen of these, they called them votarets,

who had to train for a year to learn to play the machine.

Like a musical instrument, they had to use all their fingers

and their feet with some pedals.

And it was the first time that

electricity gets applied to the problem of voice synthesis. People were really worried about

the same things that we worry about today. People wrote letters to the editor and said, you know,

could we actually end up electing an electronic president because we won't know the difference

between a human speaking and a machine speaking.

That's Sarah Bell talking about her book Vox X Machina, a cultural history of talking machines.

AI-generated voices like Siri and Alexa are now so much a part of our everyday experience

that disembodied chatbots are edging closer and closer to being accepted as real.

And it turns out it's not for the first time.

Sarah, this is really going to be fun for me because you've explored so thoroughly the history of the synthesized voice,

the computer voice that we're all so familiar with now in so many forms.

But I'm shocked to find out when synthesizing the human voice began.

It goes back to as early as, what, 1791 or 92.

Yeah, even a little bit before that, actually.

I imagine it goes back in the human imagination for even longer, but the 18th century

...