Google Clips brings AI into cameras

"How do you let yourself capture these fleeting moments, while being part of the moment?" Google asks this of us as it unveils a new camera unlike anything we’ve seen so far. Called Google Clips, it’s a self-contained camera that can take photos and videos without you prompting it to do so (but a sole button on it will let you manually record what you want). It runs pre-trained machine learning algorithms and generates clips and photos it thinks are the best ones. It then lets you export these as either photos, videos, or GIFs. Google is focusing on a specific set of users at the moment: parents and pet parents, to be specific. With the artificial intelligence component, they need to teach specific things to the machine learning model so it can start recording things like your child crawling around the living room for the first time. Google with the help of Intel’s Movidius Myriad 2 vision processing unit brings on-board AI processing to Clips.

At first glance, Clips raises some privacy concerns. But Google is quick to address this. Firstly, the clips it records are soundless. The media is all contained in the tiny teal and white cube and can only be exported to the connected smartphone. Clips comes with a 12-megapixel sensor with a 130-degree field-of-view lens capable of taking photos at 15fps. It has a 16GB internal memory and works with a handful of devices at the moment, including Pixel, Samsung Galaxy S7 and S8, and iPhone 6 and up. It’s not a cheap device, though, as it retails for US$249 (approx. CA$310).

Source: TechCrunch + The Verge