One of the biggest uses of artificial intelligence (AI) in the world right now is voice.
When embedded within electronics around the home like TVs, soundbars and lighting, voice is arguably the most intuitive way for humans to interact with technology. It’s hands free, there are no controls to learn, and you can simply “talk” to technology in the same way you talk to a human being.
Or at least that’s the idea.
Many of the voice-enabled technologies you would find around a smart home today need to connect to the internet. This enables them to reply in the appropriate manner when you talk to them. But many consumers are rightly concerned about devices in the home capturing what people are saying and transmitting it to the cloud. Many consumers are worried about the manufacturers of their devices — or worse, hackers — listening into potentially private communications. After all, if you don’t feel private in your own home, where can you feel private?
Then there are issues around connectivity and bandwidth. What if the internet goes down temporarily or you’re asking your soundbar to turn the volume up at a peak time of day when the cloud is overloaded with other people trying to do the same? Nobody wants to wait more than a couple of seconds for the device to which they’re talking to respond.
So that’s why we at XMOS want to change the way manufacturers build voice into their devices. We believe that you don’t need to connect devices to the internet to voice enable them. We now have the technology to embed intelligence within devices themselves (in an affordable way), allaying any consumer fears around unsuspecting listeners, while improving the response times and the overall experience of voice control.
Our xcore®.ai chip unlocks this potential, with prices starting at just $1 per chip. At that price point, it’s one of the only chips in the world that can embed intelligence within devices themselves, effectively giving them a brain to “think” for themselves without the need for (and the associated risks of) the internet. We explain exactly how xcore.ai delivers on the challenges associated with embedded intelligence in our Edge of Tomorrow report.
Our chip not only improves the experience of existing voice-enabled devices within the home, it can unlock new ones too. For example, how about a voice-enabled hob? If you’ve got a saucepan of rice that’s about to boil over, but your hands are dirty from cooking, you can quickly say “hob, turn down temperature” to avoid a potentially dangerous situation. For this kind of voice control to work, the device needs to be able to respond immediately every time, and therefore can’t rely on bandwidth-limited connectivity to the internet. Conversely, you might want your hob to respond only to your voice to ensure that any children within your household can’t activate the hob.
But it’s not just the home where we can voice enable. Imagine being a tourist in London and looking to use the London Underground to get around. You’re not going to get much connectivity underground, but you could go up to a voice-enabled machine and ask “Kiosk, which line do I need to take to get to Buckingham Palace?” and get an answer in seconds. That’s a much easier and more convenient experience than pouring over a map or a complicated touchscreen and trying to work it out yourself. Moreover, you’re also reducing the amount of physical contact you have with your surroundings, which in today’s Covid-19 times is of utmost importance.
Smart buildings are another area that would benefit from voice control. Many workers will be familiar with the use of building passes to gain access to buildings or certain floors or meeting rooms. But people lose passes or share them with inappropriate personnel, which damages the security of that building. But using voice biometrics instead of passes, potentially augmented by facial recognition, buildings can much more easily ensure their security.
Even simple things like coffee machines and kettles within workplaces can benefit from voice control. Items that are often touched by hundreds of people each day would feel safer and more hygienic when controlled by voice.
To summarise, the list of uses for voice control are seemingly endless. If it’s a piece of technology that humans interact with, there is usually the potential for voice control. And if there’s the potential for voice control, you’ll find xcore.ai underpinning it.
To find out more about the technical specifics of how XMOS is driving the AIoT industry, watch CEO, Mark Lippett’s recent presentation as part of Fierce AI Week, here.