Discover more from PhotoCamp Daily
Adobe has integrated its Firefly generative artificial intelligence (AI) into Photoshop, introducing a new feature called Generative Fill. This function allows the extension or removal of objects in images while maintaining the power and precision of Photoshop. Firefly, Adobe's generative AI, was announced in March and, despite not quite reaching the level of its competitors, it was developed without using artists' work to inform its algorithm. The integration of Firefly and Photoshop is described as a "deep integration", resulting in more than just image generation.
The main feature Adobe is focusing on, Generative Fill, can automatically match the perspective, lighting, and style in an image, allowing content to be extended or removed non-destructively in a few seconds using only simple text prompts. This feature enables editors to iterate rapidly on different concepts without impacting the original image, leading to boundless variations of high-quality content.
Unlike OpenAI's outpainting, Generative Fill completely reimagines the background of images, a feat Adobe demonstrates with its new feature, though with varying success. In many examples, the expanded areas of the image are not immediately noticeable, but a closer inspection reveals a difference between the original image and what was produced with Generative Fill.
Generative Fill is also available as a module in the Firefly beta, which works in a browser, enabling users to experiment with the new feature on the web. In addition to Generative Fill, Adobe previewed other Photoshop features powered by Firefly, including a "Contextual Task Bar" for intelligent subject selection and masking, a new "Remove" tool for immediate object removal, and Adjustment Presets for automatic photo editing to fit a certain look.
Adobe develops and deploys all AI capabilities according to its AI Ethics principles to ensure content and data transparency. Generative Fill supports Content Credentials, which Adobe likens to 'nutrition labels' for digital content, to ensure people know whether a piece of content was created by a human, AI-generated, or AI-edited.
Adobe's Generative Fill feature is available as a desktop-based beta and will be generally available in the second half of 2023.