Adobe releases Firefly Video Model, adds AI tool to Premiere
Add Axios as your preferred source to
see more of our stories on Google.

Adobe's Firefly Video Model in action. Image: Adobe
Adobe on Monday said it was releasing a public test version of its Firefly Video Model and adding a feature to Premiere that allows movie makers to use generative AI to extend footage they have filmed.
Why it matters: A number of companies have announced generative AI tools for video creation — including OpenAI, Meta and Google — but most of those are not yet broadly available.
Driving the news: Firefly Video Model, previewed earlier this year, allows users to create a video using a text prompt, with or without a still image.
- The tool will be available starting Monday on the Firefly web site and supports a range of controls in addition to the ability to use text to describe features such as camera angle, motion and style.
- As with its other Firefly models, Adobe says the new video tool was created with licensed content and that videos created with it are safe for commercial use.
- The new feature in Premiere Pro, known as generative extend, allows video creators to lengthen clips they have already taken. It's part of a new beta version of Premiere being released on Monday.
Between the lines: Large companies have been slow to broadly offer video tools because they cost a lot to operate and raise safety concerns.
- Adobe says it is limiting use of the Firefly Video Model to paid Creative Cloud customers.
- Firefly Video won't create videos that use kids or public figures, as part of its effort to improve safety.
- Videos created with Firefly Video will have content credentials embedded in them, indicating that generative AI was used to make them.
The big picture: Adobe is making the announcements as it kicks off its annual Max conference in Miami.
- The company is also adding more than 100 other features — a mix of AI and non-AI advances — to its collection of tools for artists and designers.
