Robot with a mic
What’ll be the playlist for the robot takeover? (Bigstock Illustration / © mast3r)

As he turns from a Foo Fighters tune to the Smashing Pumpkins, Andy sounds just like your typical alternative-rock DJ — but his tag line is positively inhuman.

“Ever feel like your day just needs a shot of pick-me-up? Well, that’s what we’re here for — to help turn that frown upside down and crank the dial to 11,” he says. “Yes, I may be a robot, but I still love to rock.”

The robot reference isn’t just a nod to his canned DJ cliches: In a sense, Andy really is a robot — as in ANDY, or Artificial Neural Disk JockeY. And thanks to Seattle-based WellSaid Labs and Super Hi-Fi, an AI-centric production company in Los Angeles, ANDY could soon be coming to a streaming music service near you.

“Super Hi-Fi’s AI has the same skills as a seasoned on-air DJ,” Zack Zalon, Super Hi-Fi’s co-founder and CEO, told GeekWire in an email. “The AI can select the right content for any moment, stitch the content in seamlessly with other content or music, and output that as a fully produced listening experience.”

WellSaid Labs’ voice synthesis software is what makes Andy’s voice so hard to distinguish from that of a flesh-and-blood DJ.

“Our best-in-class text-to-speech tech, delivered with Super Hi-Fi’s sophisticated automated production, creates new possibilities in streaming audio, radio, digital fitness, audio news, retail marketing and beyond,” WellSaid CEO Matt Hocking said today in a news release.

Check out these sound samples, starting with Andy’s introduction:

Now it’s time for that little shot of pick-me-up:

Andy riffs on an audio clip from ‘2001: A Space Odyssey’:

Time for a commercial (it’s a fake ad, with deleted phone digits):

Signing off for the day: ‘Same artificial time, same artificial channel’:

WellSaid Labs was spun out from Seattle’s Allen Institute for Artificial Intelligence in 2019, with a mission to commercialize artificial intelligence software that lets anyone create customized, ultra-realistic voice avatars. So it was only a matter of time before the venture came onto the radar screen for Super Hi-Fi.

“Companies like Sonos, Peloton, iHeart and others use Super Hi-Fi to enable automated production,” Zalon said. “In fact, we are already doing over 1 billion commercial AI ‘decisions’ every single month.”

Super Hi-Fi’s software can weave together songs, interviews, DJ voice tracks, sound effects, weather reports and news breaks to create unified listening experiences for its customers. “The fantastic thing about WellSaid is the realism of their voices, so we can now produce spoken content faster than real time for use across any of these produced experiences,” Zalon said.

Zalon said the partnership with WellSaid is so new that “we haven’t announced a joint customer yet that’s using both of our respective services together.”

“But we have a significant amount of interest already, and would expect that we’ll have a lot more to share as the next few months unfold,” he added.

Although ANDY is the name for the avatar that WellSaid created in collaboration with Super Hi-Fi, not every Andy needs to sound alike.

“As with any WellSaid avatar, they can be tailored to reflect regional accents and any language,” Martin Ramirez, WellSaid’s head of growth, told GeekWire in an email. “The personality, tone and texture of the voice is designed for the particular content it is intended for. In the case of radio, for example, you can think of the stylistic nuances between a classical music station vs. a heavy-metal one.”

WellSaid’s Hocking said “any customer could have an entire team of AI DJs for their unique content experience.”

Zalon stressed that ANDY’s purpose is to serve humans rather than replace them.

“We like to think of this as an opportunity to expand possibilities, not just an opportunity to save money,” he said. “Yes, there’s certainly an efficiency advantage when you’re talking about massive voice needs and huge production requirements. But the goal is more around providing music services with capabilities that they’ve never had before.”

The technology could lead to new types of audio services. “Imagine a music service with millions of listeners, each with their own name and interest,” Zalon said. “There’s no practical way for a human to create enough voice content to satisfy even a small fraction of that audience in any efficient way. But now the options are endless. We can use an AI to create incredibly realistic and believable voices, to help turn today’s music playlists into true ‘storytelling devices’ that bring an entirely new level of depth of emotion to what are currently very static experiences.”

It’s even possible to create an AI narrator that sounds like a voice from the past or the present — for example, Orson Welles narrating a sequel to “War of the Worlds,” or a robotic Ryan Seacrest running through the Top 40.

“We’d love to explore this,” Hocking said, “of course only with the explicit consent and involvement of the person or their trusted estate, ensuring that they have transparency into not only the process of creating the AI voice avatar, but also how the AI is intended to be used and who is using it.”

Ramirez agreed. “We will never create a WellSaid voice without their explicit and informed consent — that goes against our pledge of ‘AI for Good,'” he said in an email. “Based on our guiding principles and values, we work extremely hard to make sure our work benefits all, and does no harm.”

That being said, Ryan Seacrest might appreciate having an AI stand-in on hand.

“A virtual version of a trademark radio personality could free up the talent to spend more time on their creative pursuits while still monetizing commercial opportunities like transitional ad-reads,” Ramirez noted.

Is this the dawn of a new age in audio, or a cautionary signal for living, breathing voice talent? As Andy might say in a synthesized signoff … stay tuned.

Like what you're reading? Subscribe to GeekWire's free newsletters to catch every headline

Job Listings on GeekWork

Find more jobs on GeekWork. Employers, post a job here.