Captain Kirk is about to be very impressed.
Recently, Jabra and the London School of Economics and Political Science (LSE) shared some research about voice AI. Not to be confused with AI voices, which are getting scary good at this point, their voice AI research was focused on using talking instead of typing to interact with GenAI. The behavioral study was led by Michael Muthukrishna, a professor of economic psychology at LSE, and relied on data from 171 workers of various ages from various industries.
First, let's acknowledge that we've been working toward this level of voice interaction for some time. Dictation machines have been around since the 1880s. In more modern times, professionals have been using cassette tape-based and digital voice recorders to take notes for later transcription.
Today, it's not unusual for someone to "write" a text by speaking into their phone. Plus, I've recently reviewed products from Insta360 and viaim that literally recorded my conversations, converted them to text, and used AI to format those ramblings into useful documents. And we can't forget about Alexa or Siri, which have been doing our electronic bidding via voice for more than a decade.
The Jabra and LSE study suggests that voice will be "a primary interface for generative AI in many work scenarios" within three years.
I hope I'm making James Burke proud when I say you can follow the connections. To me, there is a clear technological line (stretching over the last 145 years or so) from recording your voice to improve workflows to using your voice to direct workflows. Now, we're poised to take the next step.
The Jabra and LSE study suggests that voice will be "a primary interface for generative AI in many work scenarios" within three years. That might seem like an aggressive timeline, but 14% of study participants already prefer voice to typing when interacting with GenAI, which exceeds the study's threshold for the early adopter phase.
In other words, voice AI is already happening—and mass adoption is right around the corner. And no, it's not just the young'uns. Older professionals were on board with voice-first AI interaction for work tasks once they used it.
Here's something that surprised me: Trust in AI rose by 33% with voice interaction compared to text. According to the report, "Voice adds something essential: a sense of connection. It makes AI feel less like a tool and more like a collaborator."
What could a voice-first future mean for integrators? Well, let's head out for a field test. Imagine you've completed a complex conference room built around AVoIP technology. There's just one problem: When you press the proverbial power button, the system doesn't work or certain devices malfunction.
Now, if it's me in that equipment room, I'm going to be yelling at clouds to express my frustration. "Why isn't this working?" I will inquire angrily to no one in particular.
But with voice AI, I could ask the system, "Why isn't this working?" and actually get a response. Then, I could tell the system to verify the connectivity of the gear, check for incompatibility between operating systems, or even perform a system-wide diagnostic. In other words, I could tell the system to fix the problem without having to type specific instructions for each level of problem solving. Nice.
Still, there's more work to do: During the study, some tasks performed via voice AI were about 20% less effective. The researchers also acknowledge the need to address privacy concerns, normalize voice AI behavior, identify the work scenarios where voice AI works best, and, of course, equip the workforce with the right technology. I'm betting Jabra has a few ideas to help you with that last one.
We're quickly approaching the promise of voice-based computer interaction that seemed so futuristic when Star Trek debuted in 1966. Of course, that show was set in the 23rd century. We seem ready to hit the USS Enterprise’s level of voice AI about 200 years ahead of schedule, thanks to the (forgive me) warp speed at which GenAI is advancing. However, before voice AI boldly goes into the mainstream, let's make sure we get it right.
