On October 15, Adobe unveiled its new AI-powered text-to-video tool, the Firefly Video Model. This innovative tool generates videos based on text prompts and sets itself apart from competitors by claiming to be fully trained on licensed content. As a result, Adobe aims to avoid the ethical and copyright challenges faced by other generative AI tools.
Adobe proudly refers to the Firefly Video Model as the "first commercially safe video model available to the public," thanks to its training on authorized materials. However, the company has not announced an official release date yet and is currently providing access only to users on a waiting list during the testing phase. Development of this new model has been ongoing since at least April 2023, with its foundational technology derived from the Firefly image synthesis model.
Similar to the text-to-image generator integrated into Photoshop, Adobe is targeting media professionals like video creators and editors with the Firefly Video Model. The company asserts that the model can produce content that seamlessly blends with traditionally produced video materials. While Adobe has not disclosed any specific clients using its video tool, reports indicate that major brands have already leveraged its image generation technology. For example, Gatorade plans to utilize Adobe's image model for custom bottle design services, and Mattel has collaborated with Adobe to design Barbie product packaging.