Recently, I had a déjà vu…

I was in my studio, recording disconnected lines for a project—long phrases, shaped by very specific emotions—and suddenly, the feeling came.

There I was, alone in front of the microphone, when an insight took me back in time. It was 2006, at the Drama School, under the direction of Silvana Garcia, working with texts by David Barthelme.

In that process, there was something very similar to what I felt now: nonlinear language, nonsense, and humanity expressed through disconnected sounds and precise emotions. It was the body—and emotion—that together gave meaning to the text. Everything could exist in the mouth—even without meaning.

And now I find myself repeating a similar logic… but this time, to train machines.

I respond to directions: “please read this line as if you were apologizing… now as if you were very grateful… now with anger, with guilt, with sadness, with joy…” It’s hard not to find it curious: artificial intelligence needs the human in order to learn even what escapes the human—the nonsensical, the noise, the in-between. Maybe because that’s exactly where we truly exist.

And it’s in that in-between—that suspended space—that the voice moves through me: between what truly expresses the human and what only tries to sound human.

Will Artificial Intelligence ever feel… or only learn how to imitate feeling?