Canva unveils AI Magic Studio and generative video with Runway

Canva, the Australian online graphic design and multimedia company whose web-based platform is used by numerous media organizations worldwide to create graphics and multimedia (including VentureBeat), is celebrating its 10th anniversary with a sweeping update of numerous AI features, including offering a new generative video tool through a partnership with AI video startup Runway ML.

With Canva’s massive Magic Studio update announced today, users of the platform can access a feature called Magic Media that allows them to simply type text into a field or upload a still image. Then, Canva’s Runway integration will generate up to 18 seconds of video based on what the user provided. In the case of text, the video will be generated based on the words provided. In the case of a still image, the image will be used as the basis of the video and motion and camera movement applied.

“We believe that AI has incredible potential to supercharge the 99% of office workers who don’t have design training or access to professional design tools,” said Cameron Adams, Canva’s co-founder and Chief Product Officer, in an exclusive video call interview with VentureBeat ahead of the announcement.

An AI-assisted design studio

But it’s not just generative video that Canva’s Magic Studio seeks to offer users.

Magic Studio is “the first all-in-one AI design platform in the market,” according to Adams, consisting of nine other major new AI features that build upon its existing AI-powered copyrighting assistant Magic Write (powered by OpenAI’s GPT-3) and text-to-image generative AI feature built atop the open source Stable Diffusion model introduced last year.

Among them are:

Magic Switch: A one-click solution for turning a single design into multiple assets, such as a presentation into an executive summary, or whiteboard of ideas into a blog post. It also supports multilingual translations.

“You can go from that document you just created over to a video that you might send to your marketing team,” Adams said, displaying in a demo to VentureBeat how a visual sales report presentation could be turned into a longer text document and then finally, transformed into a video, all within seconds.

Magic Design: Start with a text prompt or still image and then this feature lets you select color schemes and builds new imagery and designs for you to choose from.

“Any idea that you put into that search box, it can generate a unique design for,” Adams said. “You can tailor it with your brand and get it perfectly customized.”

Brand Voice: Type in guidelines and upload assets from your Brand Kit, and Canva’s AI writing and design features will follow them throughout all the other media you generate using the additional features listed here, including tone-of-voice and color schemes.

“It understands whether you want to be playful or whether your brand wants to be serious, or whether it wants to be enthusiastic or assertive or professional and it will be able to generate text appropriate to the brand itself,” Adams noted.

Text-to-Image with Multiple Styles: Now when you generate an image from text, Canva will give you many more options to choose from in different major art styles.

Magic Morph: Lets you tweak your lettering and shapes to make them shiny, puffy, 3D, or add other effects to help them stand out.

“If you want a big shiny green foil balloon, you can type that in as your prompt and it will automatically turn that shape into that exact style,” said Adams.

Magic Grab: Is like an auto-object selector an extractor that can isolate different objects within your image, similar to the functionality of Photoshop’s “Object Selection tool” or even the new object isolation feature available through Apple’s iOS Photos app.

“If you have a photo with a person in it…maybe you need a bit more space to add in text for social media posts you want to put out,” explained Adams by way of example. “Magic Grab enables this by understanding what’s in your photo, letting you grab the elements and move them across.”

Magic Expand: As the name suggests, lets you “outfill” photos and images, expanding their area beyond the original borders using AI.

New AI apps on Canva’s Apps Marketplace: The company will allow selected third-party AI companies to offer their apps through its online storefront.

Security, indemnification, and paying creators to train AI on their designs

Canva Shield is a new “enterprise-grade collection of robust safety, privacy and security controls” that allows IT administrators and design directors to encode rules for creating content in Canva through employee/team accounts, restricting certain results or content options.

Importantly for those enterprises intrigued but hesitant about AI due to underlying data copyright issues or misuses, Canva Shield offers “indemnification, providing additional peace of mind for organizations creating content with AI.”

Finally, Canva is also introducing a new “Creator Compensation Program,” a $200 million fund that will pay Canva creators “who consent to having their content used to train the company’s proprietary AI models.”

Canva also noted that it has seen massive success in its current embrace of AI, adding 65 million new monthly active users in the last year and nearly doubling its paying subscribers to 16 million.

The new Magic Studio features are now rolling out to all Canva paying subscribers.