Adobe has announced Firefly 2, the next iteration of its generative AI tool which will deliver “higher-quality” images and include a new “Generative Match” feature which seemingly spells bad news for photographers.
Firefly 2 was unveiled yesterday at the 2023 Adobe Max conference with the artificial intelligence (AI)-powered tool incorporated into Lightroom’s new lens blur feature that simulates depth of field along with a host of other tools.
However, it was the new “Generative Match” tool that will allow users to upload a reference image to guide the AI image generator to a specific style that prompted Adobe to comment that the new tools could mean less work for photographers.
Adobe is appealing to companies who want a “consistent look across assets.” It is offering brands the chance to generate hundreds, if not thousands of similar images for different uses such as websites, social media, and print advertisements.
The idea that brands can use Adobe’s tool to generate a large set of imagery could signal bad news for human creators, something Adobe’s chief technology officer for digital media, Ely Greenfield, appeared to confirm to Reuters.
“Up until a few months ago, it was still a very manual process to get all those photos — not only to take the photos but then to process them,” Greenfield says.
“Some amount of photography is going to move to virtual photography, where you’re generating from whole cloth. But a lot of it also is going to be, you do some amount of traditional photography or traditional creative work, and then you do a bunch of adaptation using this generative technology.”
Generative Match is still in beta mode which means users can’t use the output for commercial purposes yet. Any work generated by Generative Match will have a Content Credentials tag automatically attached to it which will mark it as being AI-generated.
Firefly 2 is currently available on Adobe’s web app and will be “coming soon” to its Creative Cloud apps.
Firefly 2’s New Features
At the Adobe Max conference, the company said that Firefly users have generated three billion images since the service was launched in March.
The new Firefly 2 image model will power Adobe’s popular Generative Fill tool which will be better at rendering humans with improvements to “skin, hair, eyes, hands, and body structure” and the pictures will be higher resolution and feature more vivid colors and color contrast.
Additionally, users of the Firefly web app will benefit from applying and adjusting photo settings, “similar to manual camera lens controls.”
The latest model also has an “improved understanding of text prompts and will recognize more landmarks and cultural symbols.”
Away from the world of “promptography”, there is a new Firefly vector model which creatives can use to generate “high-quality designs and illustrations.” As well as a Firefly design model which offers a “text to template” service.
Image credits: Adobe