Google is transforming all its services with AI and introducing new features to enhance user experience. At the Google I/O 2023 event, the company announced many new AI features across Google Workspace and apps, including a significant overhaul of search that puts fruitful results front and center.
It also expanded the availability of its AI chatbot Bard to everyone, putting it in the same camp as ChatGPT or Microsoft’s Bing AI. Users can now ask Bard questions about a topic, and it will provide an instant draft that they can edit or reject. The feature will be available in Gmail, Docs, and other Workspace products, and the company will also be releasing extensions that allow it to be used outside of those services.
The company also unveiled a new Magic Editor that will let you move subjects and remove unwanted elements from photos. The tool uses generative AI to replace missing objects and fill in gaps. During its I/O demo, the company showed how it could improve a shot of someone standing in front of a waterfall by moving the person toward the right of the frame and removing a visible bag strap from their shoulder.
As well as letting you reposition the subject, Magic Editor will add and extend objects such as benches to complete scenes and even fill in details that other objects or people might obscure in the photo. The company says it can do this because it has a large language model and the ability to understand the context of a photo.
Magic Editor will be included in the Google Photos app and rolled out as early access for Pixel phones later this year. Google says it will refine and improve the feature based on feedback.
The feature significantly differs from Google’s existing editing tools and will emphasize pro-level photo composition and framing. It’s similar to the sort of thing that Photoshop can do but more streamlined and easy to use.
It will also let you edit specific photo parts, such as a foreground or background object, or change things like color and brightness. But the main feature is repositioning and removing objects, which can’t be done with existing tools.
It’s a powerful tool that shows the potential of generative AI, which can create content to match the style and texture of an existing image. But as with all ML and AI, the results can vary greatly depending on how well you know what you want to do. It will be interesting to see how accurate and reliable it becomes when released to general users.