Pixel 3 uses onboard AI processing to preserve privacy and deliver new features
With privacy hot on the agenda, particularly with the recent data leak on the Google Plus network, Google has posted a blog on how the new Pixel 3 models help to protect user data and privacy by processing many of its AI functions on the device itself. Google pioneered this technology last year with the Now Playing function that brought automatic music recognition to the Pixel 2.
Google has tapped into the AI engine found in the Snapdragon 845 as well as Google-designed Pixel Visual Core incorporated into the SoC to accomplish this. One of the signature technologies carried out on board the Pixel 3 is the new Screen Call function. Onboard speech recognition is used to transcribe the dialog from the caller giving the user the opportunity to decide whether to answer the call or reject it. As this is done onboard, no aspect of the process is sent to the cloud.
Similarly, the Pixel 3 camera is also able to pull off tricks like the new Photobooth Mode. Using new low-power vision models, the new camera mode is able to analyze both facial expressions and photo quality in real time. This allows users to capture selfies and funny faces without having to reach for the shutter button. Top Shot also uses the same sort of machine learning-powered technology to suggest the best candid moments from a motion photo so you can pick from alternative HDR+ shots.
The Google Pixel 3 and Pixel 3 XL are currently available for pre-order and will ship early next month.