Ahead-looking: Google unveiled its massive push into AI on the I/O occasion this week. A lot of it revolves round image-related generative AI expertise, which the corporate is introducing for quite a few new options throughout Google Search, Android, and extra. Google is using AI to help customers in modifying photos, creating new photos, analyzing pictures, and speaking in a extra immersive method.
At Google I/O, the corporate revealed a number of initiatives leveraging generative AI imaging. Google unveiled new AI options for personalization, modifying, search, and extra.
Some Pixel telephones will obtain a brand new Google Pictures characteristic later this yr – Magic Editor. Magic Editor makes use of generative AI to allow customers to edit photos with capabilities just like highly effective instruments like Photoshop’s Content material-Conscious characteristic.
Google’s demonstration of the software program showcased somebody eradicating a strap from an individual’s physique, brightening the sky, or shifting an image’s topic. In these cases, generative AI fills within the gaps that shifting objects would go away behind on patterns reminiscent of the bottom or a park bench, concealing any indication that the image was edited.
Google can also be using generative AI in picture searches. A brand new characteristic known as “About this picture” will inform customers when Google first listed a picture, the place it initially appeared, and the place else it has been discovered on-line. This software shall be out there to US-based English customers first.
This perform goals to bolster efforts in combating misinformation ensuing from photos getting used out of context. Moreover, it can notify customers if anybody has confirmed {that a} image is AI-generated.
To show its accountability with generative AI, Google introduced that it’s going to add watermarks to all photos created with its AI instruments. Furthermore, Google Search will show generative AI watermarks from different publishers.
One space the place Google is enabling customers to make the most of AI picture technology is Android residence display personalization. Subsequent month, Pixel cellphone house owners will have the ability to reply to textual content prompts to create new wallpapers utilizing AI. On-device machine studying may even rework 2D pictures into 3D photos with the Cinematic wallpaper characteristic.
Undertaking Starline (video on high), which Google first unveiled in 2021, may even make use of AI to offer photos with further depth. Starline is a video conferencing system that goals to render name individuals in real looking 3D, creating the looks of a digital window. The preliminary prototype utilized a booth-sized array of cameras, however AI has allowed Google to develop a brand new model that generates higher photos with fewer cameras in a extra compact setup.
Google’s image-related bulletins are only a portion of its AI initiative, a necessary side of which incorporates generative AI-enhanced textual content search.