Voice Assistants and
the Internet of Skills

February 2020:

Voice Assistants And the Internet of Skills

Alexa, Cortana, Google Assistant, Siri—what’s in store for voice assistants, and how does this tech contribute to the Internet of Skills?

Voice assistants are becoming commonplace in the U.S. for many people in many situations, and their intelligence and capabilities grow with each passing year. Last year, Amazon anecdotally claimed its voice assistant technology, Alexa, was on board more than 100 million devices. Google, meanwhile, said its Google Assistant hit the 1 billion-device mark early last year.

Other competitors in the space include Apple with Siri, which the company announced in 2018 was on 500 million devices, and Microsoft, which recently added new features to its Cortana voice-assistant technology. A Microsoft spokesperson boasted about Cortana’s new productivity features, including Play My Emails, Scheduler, and Briefing email, which sends daily briefing emails to help users prepare for meetings, stay on top of tasks, and better manage their time.

In Neal Stephenson’s sci-fi book, The Diamond Age (1995), the young protagonist, Nell, gets her hands on a book that changes her life. In fact, it isn’t a book by today’s standards, but rather a high-tech interactive device called A Young Lady’s Illustrated Primer—an intelligent agent of sorts capable of reacting in realtime to its owner and its owner’s environment. The purpose of the Primer in the novel is to share skills and knowledge, essentially giving its owner information and teaching him or her whatever skills are necessary to survive and thrive.

In 2020, there’s no such thing as a Primer, but there are voice assistants, and, increasingly, there’s this concept of the growing Internet of Skills, in which technology eliminates physical barriers between people and things, enabling the sharing of knowledge and skills in realtime.

And yet, voice-assistant technology has a long way to go to reach its potential—assuming its potential is not answering relatively basic queries, automating relatively simple productivity tasks, and enabling voice control of smart-home devices like thermostats. There are cybersecurity and privacy issues to solve, and, in general, the implications of more pervasive voice-assistant technologies are not yet known.

Voice-assistant technologies complement a bigger-picture phenomenon with even more unknowns: The Internet of Skills. How will on-demand access to expertise, perhaps with a simple voice command, change the fabric of society? No one really knows the answer to this question.

Voice Assistants Are Boring

Juniper Research predicts there will be 8 billion digital voice assistants in use by 2023. According to ResearchandMarkets, the global voice-assistant market will reach $5.4 billion by 2024, up from $1.2 billion in 2018—a CAGR (compound annual growth rate) of more than 30%. And yet, Jason Hong, professor in the School of Computer Science’s Human Computer Interaction Institute at Carnegie Mellon University, says in his opinion and that of many of his colleagues, today’s voice assistants are basic. Frankly, he says, they’re also boring. “(Voice assistants are) only a thin layer over current graphical user interfaces that we have,” Hong says. “So, for example, they play music or find or add something to your calendar. It’s only saving you a few clicks on your phone or any device.”

The future of this technology, though, is less boring. Hong says there are some improvements in store for voice assistants. “Voice recognition is obviously going to get better,” he explains. “Voice assistants are also going to accommodate a lot more styles of speech too. So rather than just being very rigid in the kinds of grammar they can accept, they’re going to be a lot more flexible. And then they’re going to become much more pervasive.”

From homes to cars and maybe even bus stops, Hong and his colleagues in the Human Computer Interaction Institute at Carnegie Mellon envision all kinds of new applications for this technology as it advances. Sensing will give voice assistants a huge boost in their ability to provide contextual awareness to users.

Voice-assistant market is projected to grow from $1.2 billion in 2018 to $5.4 billion by 2024.

Source: ResearchAndMarkets
Amazon Echo

Amazon Echo

Google Home

Google Home

“You can stick a whole bunch of sensors on these devices, and they can start trying to understand things like who’s in the room, what kinds of activities people are doing, what’s the status of the room,” Hong says. “For example, if the oven is on or someone is frying something (or) the TV is on, they’ll be able to have a loT more sensing capabilities and they’ll be able to tailor responses based on that.”

Source: “Hey Alexa, What’s Up?”: Studies of In-Home Conversational Agent Usage

Future voice assistants may also provide education and learning services. Hong envisions a voice assistant acting as a sort of intelligent reading tutor, potentially listening to a child read and providing prompts or correcting pronunciation as needed.

Cosmin Munteanu, assistant professor in the Institute of Communication, Culture, Information, and Technology at the University of Toronto Mississauga, says the technology under the hood of voice assistants isn’t new; it may even be more than 40-years-old, depending on one’s perspective. And yet, the market hasn’t fully answered the question: What is it good for?

“I envision—or, I hope—that the next decade will see the design/market/business-side of this technology evolve to find its proper place in relation to user needs/wants/ability,” Munteanu says. “Although such assistants are a commercial success, I wouldn’t necessarily equate selling a $30 trinket with a major breakthrough. Turning lights on/off, asking what the weather is, controlling your music player—these are all nice, but not necessarily getting close to the potential and capabilities of such technology. I think finding the ‘sweet spot’ in terms of actual uses that make these technologies truly indispensable would be the next major evolutionary step.”

Just a few years ago, software developers would have said that having a mobile interface was nice to have, but probably not essential. Now, any enterprise product without a mobile interface is probably struggling to stay in business, according to Chris Vandersluis, president of HMS Software. “I see voice-assisted technology heading in this same direction,” he says. “We are already speaking to devices in the home, in the office, in our hand, through our headsets, and asking the devices to make an appointment, give us directions, and start a phone call. This would have been close to impossible five years ago, but the breakthroughs in voice recognition have been so successful that we don’t even think much about it anymore. That a device recognizes our voice and the command given and then attempts to fulfill that request is just part of our background now.”

As a result, more and more people are discussing how voice assistants are affecting people’s privacy and security. “It’s a two-edged sword, isn’t it,” Vandersluis asks. “The AI (artificial intelligence) required to respond to your requests does best the more it knows about you. We tend to think of a voice device as listening only when we speak to it, but of course it has to be listening for when we speak to it, and the success or failure of responding to your request has to be pumped back into the AI in order for it to learn. The problem with that is that lots and lots and lots of voice, video, and other bits of data you would have once considered private can’t be.”

Peggy's Blog

Intel Walking the Talk of 5G Network Infrastructure

The cancellation of Mobile World Congress hasn’t stopped Intel from making some pretty big noise this week.

Read More

Using Voice Assistants to Cash in on Our Data

There has been a lot of discussion as of late that advances in AI (artificial intelligence) will translate into better voice assistants.

Read More

Big Brother and Voice Assistants

Have you given any thought to the privacy and security concerns of your voice assistant or what might happen when you connect it to another smart connected device?

Read More

The Future of Voice Assistants

Artificial intelligence has truly transformed the way voice assistants are used in our daily lives, and we are only beginning to understand how they will be integrated into all of our activities in the years to come.

Read More

Voice Assistant Security and User Trust

Security for voice assistants will evolve in major ways, especially as the technology itself evolves to be able to execute more complex tasks. Sara Rampazzi, research investigator in the SPQR (Security and Privacy Research Group) Lab at the University of Michigan, says it’s quite possible that in the near future, smart-home systems with integrated voice-assistant technologies will be able to schedule medical appointments, execute house-maintenance operations, park a car in the garage, or even cook a person’s favorite meal just by recognizing her voice.

Despite all the ways the technology will enhance life and business, it’s important to consider the potential costs. “We must consider what kind of critical operations can be executed only using our voice, and what kind of consequences can happen if someone can impersonate our voice and our commands,” Rampazzi explains. “Implementing methods to verify the authenticity of the owner’s voice must be a priority to avoid the execution of critical operations or the stealing and use of sensitive information without our consent or our knowledge. Furthermore, the technology must be designed to avoid or significantly reduce the possibility of external attacks and manipulation from malicious entities.”

Security issues surrounding voice-assistant technologies, according to Carnegie Mellon’s Hong, range from the relatively gimmicky, like TV ads that intentionally activate voice assistants, to the much more insidious ideas behind things like the Dolphin Attack.

There will be 8 billion digital voice assistants in use by 2023, up from an estimated 2.5 billion at the end of 2018.

Source: Juniper Research

“(The) basic idea is if you can play some kind of audio that may be inaudible or unrecognizable to humans, like some kind of static or beep, it could activate your voice assistant and maybe even issue commands to it,” Hong explains. “It’s trying to play with the speech-recognition models behind the scenes to trick it into activating a command.”

Rampazzi and a team of academic researchers published their discovery that by using “light commands” (i.e., shining lasers at smart speakers and other devices within the laser’s line of sight), they could remotely send inaudible commands to a voice assistant. These commands could potentially include unlocking doors equipped with smart locks. In fact, as more users start linking all of their smart devices together, voice-assistant-enabled devices become a single channel for accessing a whole range of services, and that is as convenient as it is problematic.

Rampazzi’s colleague Benjamin Cyr, a University of Michigan Ph.D. student, says voice-assistant technologies are being engineered to make them seem as natural to the user as possible. They’re just one example of how technology is breaking down barriers in usability, allowing people to interact with the internet in an easier way. “But the implications of making the technology easier to use is that it may be easier for adversaries to exploit the technology,” Cyr says. “Without the proper security mindset, this technology could just be breaking down barriers that deter bad actors.”

Voice assistants that keep logs of a user’s commands can also cause potential security and privacy issues. Users must trust that these logs are being used in their best interest.

But are they? Furthermore, do users believe like these systems are eavesdropping on them? Users must also trust the technology to work in ways it’s supposed to work and not work in ways it’s not supposed to work.

“Any emerging technology requires the trust of its users,” Cyr adds. “If you don’t trust the technology to perform the task that you want, or if it performs tasks that you do not want, then you will not use it. This is doubly true in scenarios where the security of the user is at risk. At the moment, voice assistants are much more reliable than they were when they were introduced. Many people have smart-home devices that they trust to perform low-risk tasks, such as playing music, playing games, or searching the internet for knowledge. But there is still a lot of mistrust preventing wide adoption of the technology to aid in high-risk tasks, as in vehicles or making purchases. If technology makes a mistake or is exploited, it means a high cost to users. Over time, as it improves, people will trust it more. But right now, especially as problems are brought to light, it may be right for people to mistrust it.”

The Internet of Skills

User trust and security remain hurdles for voice-assistant technology, but the market projections clearly indicate growth ahead. One way the growth of voice assistants will impact society at large is by complementing and enabling the so-called Internet of Skills.

James Moar, lead analyst at Juniper Research, says: “(The Internet of Skills) will enable businesses to be more agile, enabling them to purchase remote resources to meet sudden spikes in demand. Platforms like Uber and TaskRabbit are the beginning of this shift, but in its full maturity, the Internet of Skills will enable remote working to a much larger degree than has been the case to date. However, these are unlikely to be on a large scale for jobs beyond the relatively simple, as existing hubs of expertise will continue to hold some cache in the workplace due to the physical connections made, rather than digital.”

The University of Toronto Mississauga’s Munteanu sees some obvious positives to the Internet of Skills, like opening up access to new opportunities and improving efficiency. “But I think some of these advances can lead to negative implications, which we need to be mindful of,” Munteanu warns. “For example, disruptive shifts in (the) labor force resulting not only in social consequences, but also potential long-term effects for business—e.g., trimming the workforce because such technologies allow a company to save labor costs may end up leading to problems down the road in terms of ability to stay innovative. Another implication is in widening the digital divide, or in some cases, what my research group has termed as ‘digital marginalization’—excluding many from participating in this newly created space, because we are not considering ways to ease the transition to this space.”

The Internet of Skills may enable new kinds of jobs that can’t be predicted today. It’s changing existing jobs and creating shifts in the workforce but not really eliminating any jobs … at least not yet. “What will happen when we all have these AI agents and can get just-in-time help for everything,” asks Carnegie Mellon’s Hong. “If we can get this expertise all the time, what will happen to existing jobs? It’s actually really unclear.”

Many digital natives are already accustomed to the concept of asking for help from their devices verbally, and they’re used to knowledge and, to some degree, the ability to digitally gain new skills to be accessible anywhere, anytime. Perhaps they will be the ones who push voice-assistant technology to reach its potential, hopefully taking care to address security and privacy issues along the way, and they may also be the generation that helps build an Internet of Skills.

Want to tweet about this article? Use hashtags #IoT #Internet of Things #AI #artificialintelligence #5G #cloud #machinelearning #bigdata #digitaltransformation #cybersecurity #blockchain #sustainability #futureofwork #InternetofSkills #voiceassistants

Transcription

Episode 651 | 1.21.2020

Olivier Bloch:
Simplifying the Complexities of IoT

Olivier Bloch, principal program manager, Azure IoT developer ecosystem, Microsoft, is back to talk with Peggy about the complexities with IoT (internet of Things) adoption and the biggest changes companies are facing with many vertical markets. They look into how Azure IoT is helping companies simplify IoT with IoT Plug and Play and the best things to do when adopting an IoT plan and establishing initiatives.

Guest Contributors