Text-to-video AI tools like OpenAI’s Sora are already available to users, but more competition will always benefit us in the long run. It will drive down prices and improve the quality of AI-generated videos.
Adobe is one of the companies that’s expected to innovate in this particular space, as AI-generated photos and videos fit well with its content creation products. Adobe has done exactly that in the past few years as it added more genAI features to its products.
The Adobe Firefly Video Model is Adobe’s latest AI tool. It’s a text-to-video AI app that Adobe calls “the only generative AI video model that is IP-friendly and commercially safe.”
Adobe also announced new Firefly plans that include support for generating Firefly Video. The best part is that all the Firefly content generation can happen inside a new Firefly web app, whether you’re looking to edit photos or videos with the help of Adobe’s AI models.
Firefly Video Model is available in public beta in two places. Just look for Generate Video (beta) in the Firefly web app and Generative Extend (beta) in the Firefly Video Model.
Here’s what the Firefly Video Model in Generate Video (beta) can do, according to Adobe’s announcement:
- Generate video from a text prompt or image
- Use camera angles to control shots
- Create professional-quality images from 3D Sketches
- Create atmospheric elements
- Develop custom motion design elements
The Firefly Video Model currently supports 1080p clips, but a 4K model will soon be available for pro-level workflows. Adobe also mentions a lower-resolution ideation model for high-speed iteration. The Firefly Video Model is available in the new Firefly web application, which features additional tricks beyond AI video and photo generation.
Of particular interest is Adobe’s new tool for translating audio and video in multiple languages while maintaining the original voice. This is a great way to adapt video content to multiple markets without using dubbing or subtitle services.
The samples that Adobe provided to showcase the abilities of the Firefly Video Model are just mind-blowing. The AI can create incredible clips that make it difficult to tell them apart from real videos that creators shoot in the wild.
While I am a fan of AI products and enthusiastic about what AI can do for the world, I’m always worried about AI’s ability to mislead people. Tools like Adobe’s Firefly models can be used for that.
Adobe insists in its announcement that the Firefly Video Model is the “industry’s first commercially safe generative AI video model .” Hopefully, that means the company has done extensive testing to prevent abuse. Also, I’d want the Firefly Video Model clips to be marked as created with AI clearly, though that doesn’t appear to be the case in the samples the company provided.
The more people use Adobe’s Firefly Video Model to generate videos, the more we’ll learn about it, including any safety issues. After all, the service is in public beta, so Adobe will keep monitoring it and improving it.
To take advantage of the new Adobe Firefly features, including the Firefly Video Model, you’ll need to subscribe to one of Adobe’s new Firefly plans.
Firefly Standard starts at $9.99/month. It offers unlimited access to Firefly imaging and vector features and 2,000 credits/month for video and audio generation. This gets you up to 20 5-second 1080p video generations.
Firefly Pro costs $29.99/month, providing extra access to the Firefly Video Model. You have 7,000 video/audio credits to spend per month, or up to 70 5-second 1080p video generations.
Adobe also announced a Firefly Premium plan for professionals who generate AI video content regularly. However, the company did not disclose the number of credits included in the plan or its cost.
You can get started with the Firefly web app at this link.
