SAN JOSE, CALIFORNIA: Adobe has announced the public distribution of its innovative AI model, the Firefly Video Model, designed to generate video content from text prompts. This move positions Adobe within the rapidly evolving landscape of generative artificial intelligence, aimed at transforming film and television production.
The Firefly Video Model will compete with similar offerings, including OpenAI’s Sora, released earlier this year, as well as new video tools from TikTok’s parent company, ByteDance, and Meta Platforms.
Adobe is focusing on building its models using data for which it holds usage rights, ensuring that the generated content can be legally utilized in commercial projects. While the company is currently opening access to those on its waiting list, it has not yet provided a general release date.
Although Adobe has not disclosed specific customers for its video tools, it has highlighted partnerships with notable brands. For instance, PepsiCo’s Gatorade plans to utilize Adobe’s image generation model for a platform that allows customers to order custom-designed bottles. Additionally, Mattel has been employing Adobe’s tools for packaging design for its popular Barbie dolls.
Ely Greenfield, Adobe’s Chief Technology Officer for Digital Media, emphasized the importance of practicality in the video tools. The focus is on ensuring that AI-generated footage seamlessly integrates with traditional video content. “We really focus on fine-grain control, teaching the model the concepts that video editors and videographers use things like camera position, camera angle, and camera motion.
As Adobe continues to develop its generative AI capabilities, the Firefly Video Model aims to empower everyday video creators and editors, marking a significant step forward in the creative technology space.