Speech Technology Magazine

 

The State of Artificial Intelligence

Training AI and machine learning leads to big gains and increased adoption
By Julie Knudson - Posted Mar 18, 2019
Page1 of 2 next »
Bookmark and Share

Companies are finding ways to use artificial intelligence (AI) to positively influence a huge range of tasks and processes. It’s proving effective at providing shoppers with new product recommendations, screening job candidates, and performing quality control. “PwC’s Global Artificial Intelligence Study: Exploiting the AI Revolution” estimates AI could contribute $15.7 trillion to the global economy by 2030. AI is also making inroads in speech technology, with 2018 representing a turning point for consumers, who added more voice assistants to their lives and started weaving them into their everyday routines.

The Year in Review

A few years ago, the consumer market saw a broad-scale introduction of speech technology. The voice recognition got steadily better, and so did accuracy and speed. By mid-2017, Google’s research showed that 41% of people who owned a voice-activated speaker said it felt like talking to a friend or another person.

Users were already seeing improvement in their interactions with voice assistants, but in 2018, Mark Beccue, principal analyst at Tractica, says the combination of speech technology and AI really began to permeate the landscape in a big way. “Now we’re seeing it seep into multiple parts of people’s lives, where you have people in general being much more willing to work with it.” In the earlier days, not everyone understood where AI and speech technology would come together to change how they did everyday tasks. “The idea of AI-driven speech technology is now becoming much more of a thing that people understand and want to accept,” Beccue says. Whether it’s shopping, planning a trip, or searching for information, consumers interact with AI through speech technology more than ever before.

That broader market acceptance likely got a boost in 2018 from what Tom Hebner, worldwide leader of the cognitive innovation group at Nuance, refers to as democratization of the technology. “Historically, if you wanted to have natural language understanding and speech recognition, they took a lot of compute power, and only large enterprises could afford them because it was so costly to run the technology,” he says. The mixture of wider availability, more affordable solutions, and the flexibility to run AI within a private cloud resulted in booming usage in 2018. “Now it’s more practical and we’re starting to see more applications when people get their hands on the technology and use it in their own ways,” Hebner says. 

It should come as no surprise that the organizations with a track record of investing heavily in speech technology and AI are some of the biggest innovators in the space. “Amazon and Google, I think, have improved natural language processing by leaps and bounds because they’re doing it every day,” Beccue says. “Sometimes what happens with AI and training and natural language processing is that you have to have training, and the more you can train the AI, the better it gets.” It follows that large enterprises with access to huge volumes of data can scale up their training efforts considerably and thus gain an advantage in moving new products to market.

Page1 of 2 next »