Adobe’s Firefly generative AI service is advancing as a comprehensive front-end platform for generative AI. Recently, Adobe introduced text-to-video capabilities and the option to utilize non-Adobe AI models, making these features publicly accessible.
These announcements were made during the Adobe Max conference held in London, where Firefly also inspired updates to the Creative Cloud applications. Navigating the complex terrain of AI art, Adobe has faced vocal criticism from creators about the use of machine-generated content.
To address these concerns, Adobe includes disclaimers indicating that certain images are generated by AI. Furthermore, the company is launching a beta version of a Content Authenticity app that enables creators to verify their identity and append information to their profiles on social media.
One key feature is the ability to tag content to indicate that their work should not be used to train generative AI models. Firefly has now introduced two new models: the hyper-realistic Firefly Image Model 4 and the detailed Firefly Image Model 4 Ultra.
Additionally, users can opt for non-Adobe models such as GPT for image generation and Google’s Imagen 3 and Veo 2 video models. Adobe has also added a “boards” feature, allowing users to compile various visual ideas to inspire their creations.
In a demonstration, an executive showcased how Firefly can merge two images into a cohesive output. Beyond Firefly, Photoshop is upgrading with a Composition Reference tool to guide scene arrangements and tools for selecting intricate details such as hair or facial features.
It’s also adding a new color adjustment feature. Lightroom will introduce a Select Landscape tool, while Premiere Pro rolls out its Generative Extend features, which allow users to extend video scenes with AI-generated clips and offer caption translation in multiple languages along with improved media search capabilities.