The State of Artificial Intelligence
Speech technology has been positively, and significantly, impacted by both the rapid adoption of mobile devices and the advent of voice search. As more and more people turn to Alexa, Siri, Google Home, and other applications, the need to provide them with efficient and effective interactions is critical. That’s where artificial intelligence (AI) increasingly comes into play.
With machine learning, speech-based systems can become better and more accurate over time. AI enables systems to improve as they collect more utterances, applying the data to recognize and better inform future pronunciations, speech patterns, languages, and dialects. An improved ability to grasp sentiment and context with AI is the latest hotbed of activity for technology providers.
The Year in Review
It’s no longer unusual to be in a public setting and see people “talking to themselves”—or, rather, talking to their phones and devices. AI-driven speech technology is helping to make these human-to-technology interactions more seamless and effective.
Still, says Carlos Meléndez, chief operating officer of Wovenware, a product development firm, “While everyone’s talking about AI, adoption is still in its early stages.”
Companies haven’t yet figured out how to use AI in a targeted way to solve business problems, and businesses need “to get their data out of silos and clean and prepare it to run the AI programs,” he explains.
This year RELX, a global provider of information-based analytics and decisioning tools for professional and business customers, released its second annual RELX Emerging Tech Executive Report, which highlights how executives use and view AI technologies. The report shows a jump in AI use from 48 percent in 2018 to 72 percent in 2019, with 93 percent of business leaders saying that these technologies have a very or somewhat positive impact within their industry (compared to 69 percent in 2018). AI has been used across industries, the report says, to increase efficiencies and productivity, streamline processes, and reduce costs.
It’s a time for adoption, experimentation, and innovation, says Bradley Metrock, author of More Than Just Weather and Music: 200 Ways to Use Alexa and CEO of Score Publishing. The past year did see some significant AI-related highlights, he says, noting that Google moved its powerful Duplex AI to the web, where more people can take advantage of it for conversational commerce; Amazon layered new features on top of Alexa, such as “Alexa, what am I holding?” [which Echo Show will then show you]; and Samsung introduced a system by which voice applications for its Bixby virtual assistant actually can write themselves, with initial guidelines from a creator provided.
The speech and voice recognition market is expected to be worth $26.8 billion by 2025, growing at a compound annual rate of 17.2 percent, according to a report by Meticulous Research, which also finds that AI holds the largest share of the overall voice and speech recognition market.
Technological developments in the field of speech-related AI are creating hyper-personalized customer experiences that will dominate the growth of voice-activated devices in the coming years, the research firm concluded.
Additionally, it expects widespread acceptance of AI technology, especially in education, e-commerce, healthcare, and financial services.
A Look Ahead
Adoption of AI-powered speech and voice recognition technology is, of course, just the cost of entry. In addition to adoption, says Bill Hobbib, senior vice president of marketing at DataRobot, in 2020 there will be an increased and more concerted focus on linking AI projects to key performance indicators, such as revenue growth, cost reduction, and enhanced customer experience. “Executives also will place greater emphasis on change management and encourage greater involvement from business users—rather than just data scientists or specialist teams—to increase their AI capabilities across more lines of business and processes,” he says.
The new year will also see a continued focus on making the shift from conversations to conversational AI, says Joe Petro, vice president and chief technology officer at Nuance Communications. “Conversational interfaces will begin to enable deeper interactions between human beings, with eyes and heads away from keyboards and screens and the spoken voice making a comeback,” he says. “In a doctor’s office, patients will become more literate hearing medical terms described to them that before only lived in the notes of an [electronic healthcare record]. Consumers will become educated investors listening to their financial advisers deliver more transparency than ever before.”
As conversational designers enable machines to more effectively understand and predict the intent of different users—including sarcasm—we will see better adoption than ever before.
Alex Lang, cofounder and chief operating officer of Tuya Smart, providers of a global IoT platform, agrees. “One of the ways AI technologies have advanced in recent years, and will continue to advance in the next few years,” Lang says, “is its integration with other technologies, like facial recognition or voice recognition, which unlock new capabilities for the AI and IoT industries.”
For example, he says, “We’re working to help voice recognition hear voices through more layers of sound or develop specific responses customized for each device.”
In addition, adds Meléndez, AI will both enable and destroy deepfakes—the use of deep learning and massive data to create alternative versions of a face, voice, or body in images, video, audio, or other media. While technology will allow deepfakes to gain momentum, “this technology can also be used to spot the imposters and stop them in their tracks,” he says. “In 2020, the tech giants, such as Google, Facebook, and Microsoft, will undertake larger initiatives to weed out and destroy deepfakes.”
Another area likely to see some change in 2020 related to AI is regulation, experts say.
“With the Cambridge Analytica scandal, deepfakes, and TikTok, legislators recognize the importance of regulation,” Meléndez says. In fact, he points out, “California has already made it illegal to issue deepfakes of political candidates within two months of an election. The appetite and momentum for legislation to combat bad internet actors is expected to heighten in the new year.”
AI will increasingly be applied to automatic speech recognition (ASR) in ways that will be disruptive, says John Milliken, CEO of Speechmatics. ASR, he says, is poised to disrupt many sectors, including contact centers, media and broadcasting, compliance, policing, and more. He points to some specific examples of how these impacts are likely to be applied:
• contact centers—providing real-time speech-to-text and data analytics to customer service representatives; and
• financial service providers—assuring compliance in real time with company, industry, and government standards.
In 2020, says Metrock, “expect Amazon, as well as the other competitors in the voice AI space, to amp up the functionality and features they share with the public that can be done using these voice assistants. For AI to continue to develop toward being more proactive and context-driven, and ultimately more useful for humans, these big tech companies need to succeed in driving more day-to-day utility from these assistants.”
Not yet using AI to help fuel and improve your speech technology applications? Chances are you soon will be. As Lang says: “In the future, every company will leverage AI in some fashion. Think of how companies started engaging with internet technologies as that model became the norm over the span of a few decades. The same thing will happen with AI.”
And just as we have an App Store today—and are vastly familiar and comfortable with using it—“we’ll have an AI service store in the future,” Lang says.
Linda Pophal is a freelance business journalist and content marketer who writes for various business and trade publications. Pophal does content marketing for Fortune 500 companies, small businesses, and individuals on a wide range of subjects, from human resource management and employee relations to marketing, technology, healthcare industry trends, and more.