News
Adobe Reveals New AI Tools That Will Wow Photoshop Novices
The company is forging ahead with its Firefly-based AI features, but some professionals have copyright concerns.
On Tuesday, April 23, Adobe announced a beta version of Photoshop boasting several fresh features aimed at helping users generate new images and manipulate existing files through text prompts. The latest generative AI additions harness the power of Adobe’s new Firefly Image 3 model.
Among the highlights of the update is the Generate Image tool, designed to generate images based on textual cues, providing users who struggle with a blank canvas a starting point to work from. Additionally, Generative Fill, an existing tool for background completion or image expansion, now incorporates a Reference Image function. The enhancement enables users to guide the tool’s output towards a specific aesthetic or theme by uploading an image as a reference.
The new AI tools significantly simplify the process of translating creative ideas into images and should be useful to both Photoshop novices and seasoned pros. Firefly Image 3 now has the ability to produce astonishingly realistic images, and its enhanced understanding of text prompts is now considered industry-leading.
Despite facing scrutiny over its training data and a backlash from certain segments of the creative community, Adobe remains committed to integrating generative AI features into its entire software suite. The company continues to assert that Firefly is a safe, ethical option for commercial use and positions it as an alternative to competitors like Midjourney and Stable Diffusion —both of which have faced allegations of using copyrighted material without proper attribution or compensation.
Also Read: Lebanese Newspaper Builds AI President To Beat Political Crisis
Despite Adobe’s confidence in Firefly’s abilities, recent findings have uncovered that its training dataset includes AI-generated images sourced from Midjourney and similar platforms, raising questions about the integrity of Adobe’s claims regarding the model’s commercial viability.
While Adobe maintains that its generative AI models are trained on licensed or public domain content from Adobe Stock, enthusiast Nick St. Pierre recently pointed out on X (formerly Twitter) that “over 13% of all images on Adobe Stock are AI-generated,” and that “most of the generated content comes from Dalle and Midjourney”.