Voice assistants were practically in everything at the CES trade show this year – from smart glasses and earbuds, to even smart toilets – suggesting that the future is going to be, well, smart. “Hey Google” – the command to activate Google’s digital assistant, was not only at Google’s own enormous stand, but on the Las Vegas monorail, on billboards, and on a number of other stalls at the trade show. Not to be outsmarted, Amazon was part of dozens of voice-partnership announcements at CES, with Alexa making its way into Toyota cars, Vuzix smart glasses and Kohler smart toilets.
Both Alexa and Google Assistant have opened up their Assistant voice systems to third-parties, and have already been integrated into a large number of devices in diverse markets including automotive, smart home, computers, wearables and healthcare among many others. At the event, both tech giants have confirmed industry expectation that voice assistants will play a much bigger role in consumers’ digital life this year.
Other key players in the voice ecosystem seem to have been overshadowed by Google and Amazon’s dominance.
The initial feedback from OEMs has been very positive, suggesting that voice assistant technology has been already largely adopted by end-users. Furthermore, it is now often considered as a key selling point and has driven increased usage of connected services.
At CES it was also made clear that the smartphone is no longer the unique remote controller for connected devices. This is because the graphical user interface is being challenged by voice. This shift may result in a changing relationship between the smartphone manufacturer and the customer, opening up an interesting discussion of which party owns the customer relationship. Just as OEMs took a share of control from mobile networks, voice assistants may unsettle the OEM as the one that dictates the customer experience.
What makes this revolution possible?
This voice revolution is possible because other emerging tech innovations such as the IoT. And while the intelligence is in the cloud, it is now balanced with edge computing (in this case smartphones and devices).
The other key enablers of voice technology are AI and machine learning. They are used by connected devices to interact in a more intelligent way with people and surroundings. The ability to use AI (combined with voice assistance) reinvents the customer experience and opens new business models and ecosystems. AI appears in a variety of objects from smart healthcare, automotive, smart home, services for enterprises, to name just a few.
This trend is possible thanks to the development of the neuromorphic chips, deployed in neural network and used in deep learning, virtual reality or language acquisition. The neuromorphic chips, already used in the iPhone X for the facial recognition, can mimic how the brain learns which has been demonstrated by Intel, Qualcomm and NVIDIA.
These super chipsets are already being used by new services. Canadian start-up Fluent.ai offers a multilingual command called “fluency-as-a-service”, which learns from context, behavior and speech in any language. It uses machine learning and AI to overcome the barriers that have hindered widespread adoption of voice user interface. Another is Emoshape – a microchip that gives emotional awareness to gaming, robotics, AI and IoT.
Challenges for massive adoption of voice
The future for voice assistants looks bright but it does raise a number of questions on different matters such as:
Security: A week before CES, The Register revealed two major security vulnerabilities, Spectre and Meltdown, that affect nearly every computer chip manufactured in the last 20 years. Spectre and Meltdown allow attackers to get access to data, previously considered completely protected. Voice assistants are built on large amounts of user data, so it is more important than ever to make sure they are secure locally and that the servers powering them are also as secure as possible.
Connectivity: In October, security experts discovered that all Wi-Fi networks are vulnerable to hacking. The WPA2 protocol used by vast majority of Wi-Fi connections has been broken by Belgian researchers, highlighting the potential for internet traffic to be exposed. In some scenarios, the vulnerability even leaves room for an attacker to manipulate data on a Wi-Fi network, or inject new data in, meaning that hackers could steal passwords, intercept financial data, or even manipulate commands to.
Data protection: In May 2018, the General Data Protection Regulation, which will give individuals more control over their personal data, will be applied. This means that companies should set up specific guidelines when dealing with AI to address not only the technical and data processes, but also the potential legal and ethical issues.
By : Emmanuel Legros
Head of Marketing for On-Demand Connectivity, Gemalto