Adobe debuts Firefly video model at MAX 2024, bringing AI-enhanced editing to Premiere Pro

At the Adobe MAX conference in Miami Beach, Adobe introduced the Firefly Video Model, expanding its suite of generative AI tools to video editing. Available in a limited public beta, this new model is tailored for video professionals, providing advanced capabilities for creating, editing, and enhancing video content. Adobe aims to revolutionise the creative process by introducing AI tools that save time, streamline workflows, and maintain commercial safety through licensed content.

The Firefly Video Model brings cutting-edge features like Generative Extend to Adobe Premiere Pro, allowing editors to seamlessly extend clips, fill gaps, and hold shots for precise timing. Using text or image prompts, creators can also generate video clips that include custom camera angles, motion, and zoom options. Additionally, users can transform still images into video, bringing static shots to life with Adobe’s new Image to Video capabilities.

Ashley Still, Adobe’s Senior Vice President of Digital Media, commented, “The Firefly Video Model will redefine how video professionals approach editing, enabling them to create impactful visuals more quickly and beautifully.”

Alongside the Firefly Video Model, Adobe has updated its Firefly Image Model, promising up to four times faster image generation. This enhancement is already live in applications like Photoshop and Adobe Express, offering creative professionals improved speed and new design capabilities. Similarly, Adobe has integrated Firefly into Illustrator with tools like Generative Shape Fill, which lets designers create complex patterns quickly, and Generative Recolor, allowing for unique colour transformations.

Adobe also announced advancements to other popular Creative Cloud applications, including a new 3D workspace in After Effects and expanded collaboration options in Frame.io. These upgrades aim to support more efficient project management and provide creative teams with the tools to produce professional content faster.

As AI becomes a core component of Adobe’s strategy, the company is committed to ethical AI practices. All Firefly models are trained on licensed content from Adobe Stock and public domain sources, and the company applies Content Credentials to Firefly outputs. These “digital nutrition labels” provide transparency about how the content was created, reinforcing Adobe’s standards of accountability and responsibility in AI development.

Since the initial release of Firefly in March 2023, the tools have generated over 13 billion images, gaining traction among top brands like PepsiCo, IBM, and Mattel. As part of its ongoing efforts to enhance AI’s role in creative fields, Adobe’s Firefly Video Model and other generative tools are setting new standards for efficient, AI-driven content creation.

The Firefly Video Model is currently available through a waitlist on Adobe’s website, with free access during the beta phase. Adobe plans to roll out full access and announce pricing details when the model exits beta.