How much smarter will the AI enabled smartphone become?
It is almost impossible to think of a world without the smartphone. The sheer performance this device is able to deliver in relation to its hand-sized form is a technological wonder in itself. But in recent years, functions such as immediate translation of printed text and voice, as well as recognition of music on the radio, have propelled these devices into the realm of science fiction. Such features rely on AI, but most users would probably be surprised to learn how few of them work without access to the Internet.
Despite compressing multiple processing cores into a single chip, there still isn’t enough performance available to execute the sort of AI functionality we are used to. Instead, the requests we make are sent wirelessly to cloud-based AI services. If the smartphone has no radio reception, many of these clever features become unavailable.
With the launch of the Apple iPhone 8 and iPhone X handsets, this situation has started to change. Embedded inside these phones, hidden right inside the central processor, is some dedicated neural network hardware that Apple calls a “Neural Engine”. This piece of technology is much more efficient at executing AI tasks than the CPU, which runs the iOS operating system and all that entails, or the GPU, which is responsible for the graphics display output.
Today, this enables security features, such as the 3D mapping of the users face as part of the Facial ID unlocking feature. The more entertaining Animoji feature also leverages AI, allowing the emoji image to be mapped to the user’s face in real time with minimal lag. Without AI, this feature would suffer from a poor user experience.
Other chip manufacturers, such as Samsung, Qualcomm and Huawei, are also expected to provide AI acceleration integrated alongside the processors of their next smartphone chipsets.
Next generation smartphones with AI will enable more competent virtual personal assistants (VPA), with interactions being more of a conversation, rather than just the exchange of commands we experience today. Leveraging this improved voice recognition and natural language interaction, features such as translation will become more accurate and faster. On top of this, the eco-system of wearables developing around the smartphone, such as watches, sensors and glasses, will also be able to utilize this local AI capability, allowing those devices to remain small and energy efficient.
If users are blown away by AI functionality in their smartphones today, it is difficult to image how they will react to the capability of in-handset AI in the future.