Will IoT architectures make traditional carrier networks obsolete?
From 2G, to 3G, to 4G/LTE, and now 5G, there have been marked improvements in audio quality, meaning changes for any type of speech technology that operate along these lines. However, audio operating along traditional voice channels and carrier networks could be soon replaced by data channels, with the rise of the Internet of Everything/VOIP.
As an analogy, when CDs gradually replaced vinyl, many albums were sent for remastering, as the music was no longer stored on grooves cut into vinyl, but as data. Consumers’ ears soon grew accustomed to this new quality- just as the AI systems that are a part of our everyday life will adapt to new channels. To AI systems then, this means marked improvements in speed and thresholds of accuracy, and therefore the potential uptake of more Voice-AI systems.
For example, with the concurrent rise in Conversational AI, Voice Biometric security systems are already replacing traditional, single factor passwords. These operate by analysing the sound of a person’s voice and matching this with an encrypted voiceprint.
With each deployment of new network architectures, we have seen an increase in speed through greater bandwidth. Download speed is not the only metric by which audio quality is affected, other factors such as latency, location of the caller, carrier, and the actual device being used all have their effects. 5G means that large quantities of traffic no longer need to travel along cables; data consumption spreads vast and fast across devices, from low-energy smart homes’ lightbulbs to handheld 4K mobile video streaming. With the rise of these new architectures, there are more data entry points for Voice AI to be applied.
Much of the technology surrounding new architectures could be voice controlled or at least voice aware, which is highly likely considering user-preferences and the capacity for high quality audio. This could all happen over the IP layer (e.g. over WebSockets). This is also to adapt to changing times when mobile usage and device ownership continue to grow at an exponential rate, especially in developing nations.
If the rise of conversational AI is concurrent with the IoE, then it will doubtless need easily integratable security. Voice Biometrics is the primary option for this - people are already talking to their devices. As the authentication process can be done passively - permissions for certain personal and sensitive activities, such as navigating banking, or accessing healthcare records can be linked to the voice. The new paradigm of network architecture simply facilitates conversational AI and everyday interactions with digital media transfers, everywhere.
Importance of traditional channels
This all being said, traditional phone channels may well retain an element of security simply because of the infrastructure that they operate on, familiarity with the way they’re used, and that they are familiar to older generations.
They may also remain a preferred channel of contact due to operating on what we at Aculab call “Loose Association” channels. Channels such as Voice, SMS operate on ubiquitous devices. They do not require the user to install apps to be contacted through. Second, they are channels that are the same difference between a front door and a back door - for companies to contact people, or knock, using a medium that they implicitly consent to is an extremely important way for your message to be received - it is not an invasion of privacy. In practice, this leads to users may very much wanting to be informed of planned power outages in their area, but they’d probably not be motivated to install an app from their power provider.
This is why traditional phone networks, carriers and channels will not be wholly replaced by the IoT, yet.
To read more from #DiversifyingTelecoms Campaign Week check out our landing page here.