The latest Photoshop beta includes a Generative Fill tool, which lets you add content to an image using Adobe’s Firefly AI. It’s basically a super-powered version of Photoshop’s Content-Aware Fill, and it should roll out to all users by the end of the year.
Using the Generative Fill tool is a dead-simple process. Simply select a portion of your image, type out anything you wish, and the AI does the rest. You can also let the AI use its best judgment, if you don’t feel like using a text input.
Generative Fill can add new content (such as a car) to an image. Though it can also edit the content of an image—in one example, Adobe used Generative Fill to change the texture of a sand dune. And, like Content-Aware Fill, the new Generative Fill tool can expand an image’s borders (this technique is called outpainting, and I regularly use it when creating images for Review Geek articles).
This is a major step for Photoshop, which is now the first Creative Cloud product with Firefly AI integration. I can already see how Generative Fill will speed up my workflow, though like others, I worry about misuse. Adobe promises to tag AI-generated images with identifiable metadata, but I don’t believe that metadata will stop people from spreading misinformation with generative AI.
For what it’s worth, the Firefly AI is only trained on Adobe Stock images. This ensures that AI-generated content won’t violate anyone’s copyright (as Adobe owns the license for all images in its dataset).
You can test Generative Fill by installing the Photoshop desktop beta. This tool is also available in Adobe’s Firefly web tool.