Adobe is developing a generative AI video model for its Firefly family, aimed at enhancing its Premiere Pro video editing platform. The new Firefly tools, along with potential integrations with Runway, Pika Labs, and OpenAI’s Sora models, will enable Premiere Pro users to generate video, add or remove objects using text prompts similar to Photoshop’s Generative Fill feature, and extend the length of video clips.
Unlike previous announcements related to Firefly, Adobe has not provided a release date for these new video generation tools, stating only that they will be available “this year.” The company also demonstrated the capabilities of its own video model in an early video demo, though it noted that integrating AI models from other companies like Runway, Pika, and OpenAI remains tentative.
Adobe describes these third-party AI integrations as an “early exploration” of potential future developments. The intention is to offer Premiere Pro users more options, allowing them to utilize models like Pika to extend shots or Sora and Runway AI for generating B-roll. Additionally, Adobe plans to implement Content Credentials labels on these AI-generated clips to indicate which AI models were used.
The strategy to form partnerships with third-party AI companies is seen as a logical step for Adobe. The generative capabilities of OpenAI’s Sora model have shown exceptional promise, surpassing the early previews of Adobe’s upcoming Firefly video model.
Adobe’s primary advantage lies in incorporating its Firefly models directly into its software ecosystem, which is already widely utilized by creative professionals. Integrating third-party AI models is considered essential for maintaining Adobe’s competitive edge and for enabling AI companies to reach creative professionals who prefer not to switch to a separate dedicated app.
