Improvements in natural language processing have set the stage for a revolution in how we interact with tech.

Over the past several decades, the proliferation of screens and “screen time” has been practically synonymous technology’s ever-expanding role in our lives. But this year’s CES highlights a shift in how we interact with computers: more and more, we’re bypassing screens altogether through the medium of voice.

Shawn DuBravac, chief economist of the Consumer Technology Association, said that 2017 represented an inflection point in computers’ ability to translate speech into text. When such experiments first began in 1994, he said, their error rate was about 100%. As recently as 2013, computers failed to accurately transcribe 23% of human speech.

But in 2017, they will reach parity with humans, understanding what we say at least 94% of the time. “We’re ushering in an entirely new era of faceless computing,” DuBravac said.

Apple’s Siri and Google’s Home are two of the leading “virtual assistant” brands, but CES this year showed that Amazon’s Alexa has pulled away from the pack as it adds more and more “skills,” Amazon’s term for integrations with merchants and services. “Last year we saw the very first products that integrated into Amazon’s Alexa ecosystem,” DuBravac said, and “since then we’ve seen thousands of skills released on that platform.”

Amazon Echo Dot
Amazon Echo Dot

The number of these skills grew from 1,000 in June 2016 to over 7,000 by the opening day of CES. Users can ask Alexa to add groceries to a shopping list, get a short briefing from the BBC, dim the living room lights, or look up a bank balance, to name just a few.

At the time of writing, Amazon’s Alexa skills portal counted 558 skills launched in the past seven days alone. During CES, brands from LG to Ford to Volkswagen to Huawei announced integrations with Alexa. And Amazon itself said it would connect Alexa with Amazon Restaurants, so that Amazon Prime members can use now Alexa to order takeout meals in around 20 cities.

Amazon also joined forces with Intel and the home security and automation provider MiOS to demonstrate how voice technology could work in practice. Near the Las Vegas Convention Center, the companies were giving demonstrations inside a “smart tiny home” to show some of what Alexa could do.

During our visit, representatives had to try a few times before Alexa would turn on the lights, but then she seemed to hit her stride, locking and unlocking the door, activating a pot of boiling water for morning coffee, and converting the house into a theater by drawing the blinds closed and turning on a projector.

Like any breakthrough technology, voice recognition brings its own new set of problems in addition to those that it solves. Alexa “can control anything in the home,” said Paul Lipman, CEO of BullGuard, a company focused on security for the smart home. “Imagine a scenario where your PC is hacked, and the hacker uses its speaker to say, ‘Alexa, open the garage door.’ ‘Alexa, turn off the alarm system.’ It raises the set of potential issues that are out there from a security standpoint.”

But in a year when no single gadget broke through to gain attention at CES, invisible Alexa was the star. No other trend (except perhaps autonomous driving) seemed to grab the attention of the crowd to the same extent. Marketers, advertisers and brands will be watching voice recognition closely in 2017 and beyond.