Adobe launches AI video tools, competing with OpenAI and Meta

Adobe Systems Inc. has officially begun the public rollout of its new AI model, designed to generate video content from text prompts. This innovative technology, named the Firefly Video Model, aims to transform film and television production by leveraging generative artificial intelligence. Adobe enters a competitive landscape, facing off against OpenAI's recently launched Sora, as well as video tools introduced by TikTok's parent company, ByteDance, and Meta Platforms.

In an effort to compete effectively, Adobe is focusing on developing models trained specifically on data it has the rights to utilize, ensuring that their outputs can be legally used in commercial projects. Based in San Jose, California, Adobe plans to gradually release the tool to users who have signed up for its waiting list, although a general release date has yet to be announced.

While no specific customers have been disclosed for the new video tools, Adobe revealed that Gatorade, owned by PepsiCo, will utilize its image generation model for a platform that allows customers to create custom bottle designs. Additionally, Mattel has been employing Adobe's tools for the packaging design of its popular Barbie dolls.

Ely Greenfield, Adobe's chief technology officer for digital media, emphasized the company's commitment to making its video tools accessible and practical for everyday creators and editors. The focus is on ensuring that the generated footage seamlessly integrates with traditional video content. Greenfield noted the importance of fine-grain control, teaching the model essential concepts used by video editors and videographers, such as camera angles, positions, and motions. This approach aims to empower creators and enhance the overall production process.