Adobe is preparing to launch its generative AI video model, Firefly, later this year, following the success of its tools for still images. The company aims to provide professional creators with a "commercially safe" AI option by training Firefly exclusively on licensed content. This ensures that users can confidently incorporate AI-generated elements into their projects without concerns about intellectual property issues.
The Firefly video model will support text-to-video and image-to-video features, enabling editors to add new elements to footage or enhance existing projects with cinematic effects. Adobe will also offer a variety of controls, allowing users to adjust critical aspects like camera angles and motion for greater creative flexibility. Initially, Firefly will be available on Adobe's Firefly website, and the company plans to embed its "Generative Extend" feature directly into Premiere Pro for seamless integration with existing workflows.
Ely Greenfield, Adobe’s chief technology officer for digital media, emphasized that Firefly respects creators’ rights, highlighting the company’s commitment to only using authorized content for training its models. This approach sets Adobe apart from other generative AI video tools, which may not provide the same level of content safety. Firefly aims to give video editors more control and efficiency, enhancing their ability to produce innovative content.