When you speak to an Amazon Echo device, so far in its history, Alexa would connect to the internet to process requests and deliver a response. Of course, this has sparked concerns among some consumers, since it might feel worrisome to have a Wi-Fi-connected smart device listening for you to say a wake word at all times. But Amazon announced at its fall event today that new Amazon Echo devices, starting with the Echo Show 10 and the latest-generation Echo, will be able to process voice commands locally, rather than sending information to the cloud.
Amazon claims it's the first company to offer this kind of privacy-first technology on smart speakers. Last year, Amazon unveiled the AZ1 Neural Edge processor, which powers current Amazon Echo devices, but its new Echo Show 15 will be powered by the AZ2 processor. Amazon said that this processor can do 22 times more TOPS (trillions of operations per second) than the previous generation.
In addition to local voice processing, the Echo Show 15 will support a new feature called visual ID. Alexa will be able to recognize individual users when they are within the frame of view. This means that it can deliver personalized content, knowing to display your calendar as opposed to another household member's calendar, for example.
Amazon said that visual ID is also built with privacy at its foundation. The feature is optional, and users must enroll to use it -- all processing happens locally on the device, and users can delete their visual ID profile at any time.