Skip to main content
“Tonight, we picture life as an intricate tapestry, where history weaves our collective identity, offering lessons and grounding us,” says an unidentified voice on Saturday, Feb. 1, introducing KBOO.FM’s monthly late-night show Parsing Sound. Feminine and vaguely British, the voice belongs to Angel Aura, an artificial intelligence voice model created by Jamaal Hale, a Parsing Sound co-founder and owner of the graphic design firm Good Green.
Jamaal’s brother Malik Hale, one of Parsing Sound’s four co-founders, tells WW that Angel Aura has been used on their show since its 2020 launch during the pandemic. Dissatisfied with the sound of his own voice and unable to connect with anyone else during lockdown, Hale used Angel Aura to introduce each episode and fulfill certain KBOO duties like traffic reports. Its script is written by the Parsing Sound team, which thematically curates its playlist. While Angel Aura has interacted with some show guests, the voice model is incapable of spontaneous dialogue.
“Instead of looking up the dictionary and looking up the answer, you can speak and say, ‘Hey, what does this word mean?’ and it will verbally tell you right back,” Hale says. “But there’s no dialogue going back and forth.”
Hale is familiar with arguments that AI will replace human jobs—as arose in 2023 when Live 95.5 FM introduced its own automated voice to read traffic reports—but says the technology is nowhere near ready to replace professionals. He says Angel Aura is similar in practice to TikTok’s vocal text reader, but requires a technician’s fine tuning to make its delivery sound more naturally human.
“You have to have a human actually running it at all times to get the best out of it,” Hale says. “As far as taking people’s jobs, we are light years away from that.”
KBOO, the community radio station which largely relies on volunteers, announced layoffs last fall amid an ongoing budget deficit, though none of KBOO’s layoffs came from its talent pool. KBOO station manager Nathan Vandiver tells WW he was unaware that Parsing Sound uses AI, and says the station has no formal policy regulating its use.
“While we have basic ground rules and guidelines for our programmers and try to fit their programs on the schedule in a time that makes sense for their genre, KBOO programmers enjoy great latitude for creativity, so we don’t always know what’s on their show ahead of time,” he says via email. “I think some of our programmers have strong feelings about it, but I’m not sure how a policy would shake out for KBOO yet. Also, if we came across an issue, we would likely adjust practice before the policy was finalized.”
Hale points to other practical applications for voice models like Angel Aura.
“We started to hear from the music community about people who can hear but can’t really talk, so they say this is awesome,” he says. “There’s classes where people teach using text-to-voice…but there are people who think teachers will be replaced by AI and that’s not possible. You might be able to program something with a curriculum, but how are you going to give those kids that extra one-on-one help, or what if the kids aren’t tech-savvy? Those are just fears more than anything else.”