Starting today, a first powerful Firefly feature is available to try out in the latest beta version of Photoshop. "Generative Fill" is based on a generative AI model that can conclusively fill in masked areas in an image. As with Stable Diffusion, not only can objects be deleted or changed, but an image can also be extended in its frame - by the Firefly AI "spinning on" the presented image motif:
While none of this is new to the AI scene, this is likely to be the dam-breaker for generative AI applications - because now millions of Photoshop users have direct access to features that have often been dismissed as toys for nerds in recent months.
But with Generative Fill, using potent AI tools in Photoshop has now become "AI-nder easy," as the following tutorial demonstrates:
You can literally tell how much the resulting motifs still surprise the presenter. And indeed, "Generative Fill" is probably the most powerful new Photoshop function of recent years. Professional image manipulation now only requires a selection of the area to be changed and, if necessary, a few fitting words.
Anyone who is still not convinced after this tutorial that AI will turn the entire media industry upside down has not heard the shot. And Adobe is likely to give the majority of its Photoshop users sleepless nights with this step. But in a positive sense, because once you've touched the new tool, it's hard to let go.
In addition, Adobe can also offer its users a licensing model that should make the commercial exploitation of their own works much more legally secure - because Adobe has all the rights to the images that have been incorporated into the AI model during training. With Stable Diffusion or Midjourney, on the other hand, the legal situation is still largely unresolved.