Key Points
- Adobe is launching Adobe Firefly Video Model, an AI-powered video creation and editing tool, later this year.
- The tool generates five-second clips from prompts, supports text and image inputs, and allows customization of camera angles, motion, and zoom.
- The AI model is trained on Adobe Stock content, ensuring it does not include intellectual property, trademarks, or customer content.
- Adobe aims to compete with AI video tools by focusing on quality and prompt accuracy tailored to videographer needs.
Adobe is set to release a new generative AI-powered video creation and editing tool, Adobe Firefly Video Model, in a limited beta version later this year. The new tool aims to enhance Adobe’s offerings for creative professionals, joining the company’s existing suite of Firefly applications known for generating still images, designs, and vector graphics.
With the launch of Firefly Video Model, Adobe seeks to establish a foothold in the rapidly growing market of AI-based video generation tools, which is currently being explored by competitors like OpenAI’s Sora, Stability AI’s Stable Video Diffusion, and other AI video applications from smaller startups. The new tool will allow users to generate five-second video clips from a single prompt, supporting text and image inputs. Additionally, users can customize camera angles, panning, motion, and zoom to better align with their creative vision.
“We’ve invested in making this model reach the level of quality and prompt understanding that videographers expect,” said Alexandru Costin, Adobe’s Vice President of Generative AI. “Our focus has been on ensuring the tool respects the guidance from videographers, offering a higher level of prompt accuracy than other AI video models currently available.”
Adobe emphasized that the Firefly Video Model is trained exclusively on content the company has the right to use, avoiding any Adobe customer content. The training data includes content from Adobe Stock, a database comprising 400 million images, illustrations, and videos, all curated to exclude intellectual property, trademarks, or recognizable characters.
Alongside the new video model, Adobe is also introducing Generative Extend, a feature for its Premiere Pro video editing software. Generative Extend can automatically add up to two seconds of footage by generating suitable inserts to fill gaps in video clips. First previewed in April, the tool has received strong positive feedback from users, according to Costin.
By expanding its AI capabilities in video creation, Adobe aims to strengthen its position in the creative technology market and provide innovative tools that meet the evolving needs of video professionals.
