MIAMI — Adobe unveiled artificial intelligence tools that can create and modify videos, joining Big Tech companies and startups in trying to capitalise on demand for the emerging technology.
One feature, integrated into Adobe’s video-editing software, Premiere, will let users extend video clips using generative AI, the company announced at its annual product conference in Miami.
This move puts Adobe in competition with other major players like OpenAI’s Sora, ByteDance, and Meta Platforms, all of which have recently introduced similar video tools.
Unlike its larger competitors, Adobe focuses on using data it has legal rights, ensuring that the AI-generated content can be used commercially.
The company is gradually rolling out access to Firefly Video for users on its waiting list, though a full release date has not been provided.
Dubbed Adobe Firefly Video Model, the artificial intelligence tool will be released in beta and will join the Photoshop maker’s existing line of Firefly image-generating applications that allow users to produce still images, designs and vector graphics.
The model will establish Adobe in the growing market for AI-based video generation tools, a space already targeted by OpenAI’s Sora, Stability AI’s Stable Video Diffusion and other AI video apps from smaller startups.
The tool can generate a five-second clip for a single prompt and can interpret both text and image prompts, said Alexandru Costin, vice president of generative AI at Adobe. Users can also specify the required camera angle, panning, motion and zoom.
TikTok to label AI-generated images, video from OpenAI and elsewhere
“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,” Costin told Reuters in an interview.
Adobe said the video model is trained on public domain or licensed content that it has permission to use, and not on any Adobe customer content.
“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,” Costin said.
Adobe is also rolling out Generative Extend, a tool that will be available in its Premiere Pro video editing software, which can extend any existing clip by two seconds by generating an appropriate insert to fill gaps in the footage.
Discussion about this post