Adobe announced the upcoming beta release of its new AI-powered video creation and editing tool, Adobe Firefly Video Model, which aims to expand its offerings to creative professionals. Set for a limited release later this year, this tool will complement Adobe’s existing Firefly applications for generating still images, designs, and vector graphics.
The Adobe Firefly Video Model positions Adobe in the competitive AI-based video generation market, which includes OpenAI’s Sora, Stability AI’s Stable Video Diffusion, and other applications from emerging startups. According to Alexandru Costin, Adobe’s vice president of Generative AI, this model can generate a five-second video clip from a single prompt and interpret both text and image inputs, allowing users to specify details like camera angles and motion effects.
In an interview with Reuters, Costin emphasized the company’s commitment to achieving high-quality output and accurate, prompt interpretation to meet videographers’ expectations. He noted that the model uses data from Adobe Stock, which includes over 400 million images, illustrations, and videos that are carefully curated to avoid intellectual property issues.
Additionally, Adobe is introducing Generative Extend, a feature within its Premiere Pro software that can automatically extend video clips by two seconds, seamlessly filling gaps in footage. This tool, first previewed in April, has already received overwhelmingly positive user feedback.
