As competition in the AI industry intensifies, Adobe is positioning itself to stay ahead with the launch of its new Firefly Video Model. As technology companies vie to outdo one another, Adobe’s latest AI tool integrates across its creative products, including Premiere Pro, enabling users to extend footage and generate video from both images and text prompts. In essence, Firefly for video works similarly to ChatGPT, but for video creation.
Adobe Premiere Pro introduces Generative Extend
The first tool introduced under the Firefly Video Model is the Generative Extend feature, launched in beta for Adobe Premiere Pro. This tool allows users to extend video footage and make mid-shot adjustments as needed. However, the feature comes with some limitations. Currently, video clips can only be extended by two seconds, underscoring that this AI tool is still in its early stages and primarily aimed at facilitating minor tweaks.
Extended clips generated by Firefly are limited to either 720p or 1080p at 24 frames per second. In addition to video, Adobe’s AI can handle audio, but only for short, 10-second clips, and it does not support dialogue or music at this time.
Adobe expands its AI offering
Adobe appears fully committed to enhancing its AI capabilities, as it has released two video generation tools and is also testing a web app version of Firefly in a limited public beta. The AI can generate various styles of video, including real-style frames, 3D animation, and stop-motion, with customizable camera angles and shooting distances.
Despite these advancements, Firefly currently has some limitations that prevent it from replacing long-form video creation. At present, the tool can only generate five-second video clips at a maximum resolution of 720p and 24 frames per second. In comparison, OpenAI’s video generation tool, Sora, can create videos up to one minute long, although it is not yet available to the public.
Adobe has emphasized that its AI models have been trained only on licensed content, setting it apart from other generative AI systems that have faced criticism for using unapproved material.
Adobe Max conference reveals AI-powered features
At Adobe’s Max conference, the company announced a broad range of AI-powered upgrades across its suite of products, including Photoshop, Illustrator, InDesign, and Premiere Pro. These enhancements promise to reshape how users interact with the Adobe Creative Cloud, with Firefly AI at the center of these changes.
One notable update is in Photoshop, which now includes an AI-powered distraction removal tool within the existing Remove tool. This feature identifies and eliminates visual distractions—such as people, wires, and branches—with a single click, improving image clarity and composition. Photoshop has already integrated several AI features, including Generative Fill, Generative Expand, Generate Similar, and Generate Background, which had been in beta but are now permanently available.
Firefly 3 model enhancements
The Firefly Image 3 AI Model is another significant advancement, offering an expanded variety of generated images with enhanced realism. This new model also improves the ability to interpret complex prompts, a capability that will be rolled out across the Adobe Suite as Firefly becomes more deeply embedded in these tools.
Illustrator will also benefit from AI upgrades, including a new “Objects on Path” feature, which gives users more control when aligning objects along a designated path. This update will make arranging and moving objects more efficient. Additionally, Illustrator’s core Image Trace feature will be enhanced with AI to improve precision and flexibility.
Premiere Pro’s AI video tools
Premiere Pro, one of Adobe’s flagship video editing tools, will see significant upgrades as well. The Firefly Video AI Model has been trained using Adobe Stock’s extensive library and will continue to evolve through user feedback during the beta phase. The Generative Extend tool will allow users to extend videos by generating new frames based on existing footage. The AI analyzes the video to ensure the newly generated frames seamlessly blend with the surrounding content, offering a smooth extension.