Adobe released Firefly Image 3, the third generation of its AI-powered image generator, on Tuesday. The company claims this model understands prompts better, can provide greater detail and has “better lighting, positioning, attention to detail [and] advancements to text display.”
What it is. Adobe Firefly is a generative AI tool for creating and modifying images, illustrations, and fonts. It is available in Photoshop (beta) and Adobe’s Firefly web app. The company says a video-generation version will be available in the Premier Pro video editing tool later this year.
Dig deeper: Legal risks loom for Firefly users after Adobe’s AI image tool training exposed
New in Photoshop. Adobe has added tools to make adding or tweaking elements of images easier. These include the Generate Image feature (below), which lets users generate images from text prompts in the application.
Generative Fill lets users generate new elements based on uploaded reference images (below).
Generate Similiar can create variations of imported images (below) and add them to projects.
The fill and generate similar tools will be particularly useful for matching brand styles and themes in creative assets.
Why we care. Better image-generation tools make work easier for marketers. We are unabashedly and 100% in favor of anything that does this.
Let the user beware. A recent study found that, despite Adobe’s claims, Firefly wasn’t trained entirely on images licensed by the company — raising issues of legal liability. The company has said it will indemnify enterprise users sued over using its imagery. However, that protection goes away if users’ reference images introduce copyright infringement.
