Intel’s offline voice recognition tech could be a Siri-killer

Intel is about to bring voice recognition technology full circle.

While voice recognition has been available on desktops for over a decade, it’s found a new life on mobile devices, where it powers things like Apple’s Siri virtual assistant and Google Now. But to make that platform leap, tech companies have had to rely on cloud processing, which adds significant delays and makes voice commands unavailable without an Internet connection.

Intel’s big fix? Bring voice recognition back to the device.

Together with an unnamed voice recognition technology provider, Intel has developed a new solution for its latest mobile processors to power voice commands locally, foregoing a trip to the cloud, reports Quartz’s Christopher Mims. That could ultimately lead to more responsive voice commands and the ability to do things like dictate emails even when you’re in the middle of nowhere.

“It’s fine if [voice recognition systems] can’t make a dinner reservation because the phone can’t get to the cloud,” Intel’s head of wearables, Mike Bell, tells Quartz. “But why can’t it get me Google Maps on the phone or turn off the volume?”

Intel will first implement the technology in its “Jarvis” headset, a device announced earlier this month that will serve as the gateway to the company’s own virtual assistant.

Intel formed an agreement with Nuance, creator of the popular Dragon Naturally Speaking voice recognition software, two years ago to bring its technology to Ultrabooks. Earlier this month, the two companies announced the first round of computers to receive that tech. Given this partnership, as well as Nuance’s dominant role in the voice recognition space (it also powers Siri), I wouldn’t be surprised if Nuance is Intel’s secret partner for this new mobile voice recognition tech.