Look out, Canva: Google Pics’ AI editing is scary good
Summary created by Smart Answers AI
In summary:
- Google launched Pics, a new AI-powered image editing app using Nano Banana 2 generative AI that can generate, edit, and manipulate images similar to Canva.
- PCWorld notes the app excels at AI-driven text manipulation within images, avoiding font approximation issues that plague Canva’s Magic Layers feature.
- Currently in limited testing, Pics will eventually integrate into Google Workspace, though long-term success depends on Google’s commitment and subscription pricing model.
In March, Canva made a big deal of Magic Layers, and its ability to treat different elements of an image as editable layers. On first blush, Google’s new Google Pics app looks like it can do the same — and then some.
Google Pics is a new, standalone Google app, powered by Google’s Nano Banana 2 generative AI engine, that Google launched here at its Google I/O conference in Mountain View, Calif. It’s being tested by a limited group of testers. Eventually, however, Google plans to make it part of its Workspace suite that includes Sheets, Docs, and Slides.
Pics can be used to generate, edit, and otherwise manipulate images, much like Canva — although Canva’s ground-up design elements and third-party integrations (even into Google!) — give it a substantial advantage.
Still, Pics is impressive, at least within the narrow confines of Google’s demo suite here at Google I/O. Like many productivity apps, it feels decidedly iterative: Pics is an app that can both create and edit using generative AI, much like Google’s photo editing tools on its Android platform.
One notable feature of Pics is how well it edits and manipulates text, working within the abilities of AI rather than defined fonts.
Put another way, a user applying Canva’s Magic Layers can extract text from an image and Canva will try to map it to a font it understands, at least in my experience using the tool. This can work flawlessly, especially when using a known font. But when Canva encounters a font it doesn’t understand, it has to approximate it. In that scenario, the product looks a little off.
Pics uses AI and just AI, which works effectively. Consider how well a modern generative AI image model approximates photography — gone are the days of blobby images with multiple fingers. Editing a fake promotional flyer required a simple click, a few changes in a text box, and a wait of about 10 seconds to recalculate the image. (Google representatives said that the time and efficiency of the model will further improve as users use and train it.)
By now, Canva’s integrated design toolset is a polished machine. Google is well, Google, with a reputation for launching ambitious projects and then unceremoniously killing them — like its Sora AI video generator, for example. Workspace, however, seems to have an aura of permanence about it, possibly because it falls under the umbrella of a subscription. In any event, Pics looks like a legitimate tool you’ll want to use…though you’ll need to pony up for a subscription, too.
