Snapchat announced at its annual Snap Partner Summit on Tuesday that it's introducing a new AI video generation tool for creators that will allow select creators to generate AI videos from text prompts, and soon from image prompts as well.
Snap said the tool is available in beta on the web to some creators starting today, and that there are currently no plans to open Snap AI Video to everyone outside of creators.
The company didn't reveal any details about the AI video generation tool during the keynote, but a company spokesperson told TechCrunch in an email that the tool is based on Snap's own underlying video model.
“We know it's important to let our community know when AI is used to generate content,” Snap's vice president of product, Cesi Mourkogiannis, said during the keynote, “that's why we use icons and contextual cards to let people know that a Snap was made with Snap AI.”
Image credit: Snap
All content created with Snap AI Video contains a watermark that will be visible when the content is downloaded and shared.
A Snap spokesperson said its models undergo safety assessments and testing to prevent the creation of harmful content, and that the company puts in place additional safeguards during the model training process, when creating prompts and at the output stage.
Snap's announcement puts the company one step ahead of social media peers like TikTok and Instagram, which have yet to release AI text-to-video generation tools. Meta announced a text-to-video tool called “Make-A-Video” in 2022 but has not yet made it publicly available.
The company appears to want to compete with startups like Runway, as well as popular companies like OpenAI and Adobe, which plans to publicly release their text-to-video AI generator later this year.
However, Snap hasn't released any examples of the output of its AI video generation tool, so it's unclear how its product compares to others.