Adobe today announced that a new ‘Generative Fill’ feature is coming to its ubiquitous photo editing software later this year.
The company promises a “magical new way to work” as the Firefly-powered feature lets you add, remove and expand visual content based on natural language text prompts.
“Generative Fill combines the speed and ease of generative AI with the power and precision of Photoshop, enabling customers to bring their vision to life at the speed of their imagination,” said Ashley Still, Adobe’s Senior Vice President of Digital Media.
Adobe’s Generative Fill is described as the equivalent of DALL-E 2’s inpainting (AI-generated content within a portion of the image) and outpainting (AI-generated content that extends beyond the boundaries of the image).
For example, if you want to paint the sky to make it look surreal in a photo you’ve taken, select that area and type something like “surreal sky with weird colors” in the prompt field.
Or if you’ve taken a photo that you want to have a wider aspect ratio, you can select the area outside of it and ask it to stretch the scene.
Invisible Digital Signature For Image Information
To help differentiate its AI work on an ethical level, the company says its current generation model only learns from Adobe Stock images and “other public domain content without copyright restrictions.”
Also, as part of Adobe’s Content Credentials initiative, AI images made in Photoshop will be encoded with an invisible digital signature that indicates whether they are man-made or AI.
As generative AI makes it increasingly difficult to separate the organic from the algorithmic, and artists worry about their work being plagiarized by career-killing machines, Adobe says it is taking a more transparent approach.