
I was getting my dopamine rush the other day scrolling through some YouTube Shorts when the narrator clearly pronounced the word “boundaries” as “boun-dairies,” with the first syllable sounding like the word for a person, place or thing but beginning with “b” and the second sounding like the place where milk cows work.
It’s a pretty safe bet the narration was AI-generated with a text-to-speech algorithm. But calling attention to the mispronunciation isn’t intended as a gotcha aimed at the technology or the creator of the content. Rather, it’s a sign of just how good this technology has become. If it were not for the exception my ear registered in the form of “dairies,” I would not have thought twice about the authenticity of the voice.