Adobe Will Soon Roll Out Generative AI Video Creation Tool

The tool can translate the text and visual clues and can create a five-second video.
Ayesha Anwar
By Ayesha Anwar
6 Min Read

Adobe will launch a new generative AI-powered video creation and editing tool in a limited release later this year. The company released this statement on Wednesday in an effort to improve the applications that it provides to creative workers.

The latest addition to the enormous design software company’s portfolio of generative AI services, known as Adobe Firefly, is the AI tool known as Adobe Firefly Video Model.

The new AI tool with text-to-video and image-to-video abilities will only be launched in beta later this year and will join Photoshop Maker’s present line of Firefly image-generating applications. Adobe also announced that they are improving its Premiere Pro software for video editing with Generative Extend, a feature that stretches the canvas of a video or image frame with recently produced backdrops.

The company said in a statement, “The Firefly Video Model is designed to be commercially safe and is only trained on content we have permission to use—never on Adobe customer content.” 

Adobe will take a more prominent place in the expanding market for AI-based creation tools with this approach, which is already being explored by smaller businesses, for example, Sora from OpenAI, Stable Video Diffusion from Stability AI, and others.

According to the Vice President of Generative AI at Adobe, the tool can translate the text as well as visual clues and can also create a five-second video clip for a single search. In addition, users may choose the essential motion, zoom, and camera angle.

Reuters were told by Costin in an interview, “We’ve invested in making this model reach the level of quality and prompt understanding that videographers expect. We’ve invested in making sure we really pay attention to the prompt… respecting guidance from videographers much better than other (AI video) models.”

Adobe also stated that the video model was trained on licensed or public domain content that it was given permission to use rather than on content owned by any customer of Adobe.

Costin gave a statement, “We only train them on the Adobe Stock database of content that contains 400 million images, illustrations, and videos that are curated to not contain intellectual property, trademarks, or recognizable characters.” 

According to Costin, the tool was first previewed in April and since then has seen “a huge positive reaction from all of our customers.”

TAGGED: ,
Share This Article
Leave a comment